{"Bibliographic":{"Title":"Atmospheric Transport and Diffusion in the Planetary Boundary Layer July 1971-June 1972","Authors":"","Publication date":"1973","Publisher":""},"Administrative":{"Date created":"08-17-2023","Language":"English","Rights":"CC 0","Size":"0000159762"},"Pages":["NOAA TM ERL ARL-39\nA UNITED STATES\nDEPARTMENT OF\nCOMMERCE\nNOAA Technical Memorandum ERL ARL-39\nPUBLICATION\nDEPARTMENT OF and\nU.S. DEPARTMENT OF COMMERCE\nNATIONAL OCEANIC AND ATMOSPHERIC ADMINISTRATION\n.\nEnvironmental Research Laboratories\nSTATES OF\nERL\nARL 39\nAtmospheric Transport and Diffusion\nin the Planetary Boundary Layer\nJuly 1971-June 1972\nAir Resources\nLaboratories\nSILVER SPRING,\nMARYLAND\nJune 1973","ENVIRONMENTAL RESEARCH LABORATORIES\nAIR RESOURCES LABORATORIES\nOF\nwhite COMMUNITY\nSTATES OF\nIMPORTANT NOTICE\nTechnical Memoranda are used to insure prompt dissemi-\nnation of special studies which, though of interest to\nthe scientific community, may not be ready for formal\npublication. Since these papers may later be published\nin a modified form to include more recent information\nor research results, abstracting, citing, or reproduc-\ning this paper in the open literature is not encouraged.\nContact the author for additional information on the\nsubject matter discussed in this Memorandum.\nNATIONAL OCEANIC AND ATMOSPHERIC ADMINISTRATION","Distining\nERL ARL-39\nU.S. DEPARTMENT OF COMMERCE\nNational Oceanic and Atmospheric Administration\nEnvironmental Research Laboratories\nNOAA Technical Memorandum ERL ARL-39\nATMOSPHERIC TRANSPORT AND DIFFUSION\nIN THE PLANETARY BOUNDARY LAYER\nJULY 1971 - JUNE 1972\nI. Van der Hoven, Editor\nContributors\nJ. K. Angell\nG. A. Herbert\nA. B. Bernstein\nJ. F. Sagendorf\nD. J. Bjorem\nG. E. Start\nH. L. Boen\nL. L. Wendell\nC. R. Dickson\nAnnual Research Program Review\nJuly 1971 - June 1972\nfor\nU. S. Atomic Energy Commission\nOF\nCOMMUNITY\nAir Resources Laboratories\nSilver Spring, Maryland\nANGELES\nwith\nJune 1973\nSTATES OF","DISCLAIMER\nThe Environmental Research Laboratories, National\nOceanic and Atmospheric Administration, U.S. Dept.\nof Commerce, does not approve, recommend or endorse\nany proprietary product or proprietary material\nmentioned in this publication. No reference shall\nbe made to the Environmental Research Laboratories,\nor to this publication furnished by the Environmental\nResearch Laboratories, in any advertising or sales\npromotion which would indicate or imply that the\nEnvironmental Research Laboratories approves, recom-\nmends or endorses any proprietary product or proprie-\ntary material mentioned herein, or which has as its\npurpose an intent to cause directly or indirectly the\nadvertised product to be used or purchased because of\nthis Environmental Research Laboratories publication.\nThe findings of this report are not to be construed\nas an official Department of Commerce position, unless\nso designated by other authorized documents.\nii","TABLE OF CONTENTS\nPage\niv\nLIST OF FIGURES\nvii\nLIST OF TABLES\n1. URBAN PLANETARY BOUNDARY LAYER STUDIES\n(OKLAHOMA CITY)\n1\n1\n1.1 Introduction\n2\n1.2 The WKY Tower Measurements\n1.3 Three-Dimensional Tetroon Trajectories\n2\n1.4\nPibal - Radiosonde Measurements\n10\n1.5 Fast-Response Wind and Temperature\n10\nFluctuation Measurements\n10\n1.5.1\nObjective\n11\n1.5.2\nInstrumentation\n12\n1.5.3\nExposure\nSignal Conditioning, Recording,\n1.5.4\n14\nand Calibrating\n1.6 Mesoscale Windfield and Transport\n15\nAnalysis\n2. PRELIMINARY WORK ON HASWELL, COLORADO,\n27\nMESOSCALE STUDY\n3. CONVERSION OF THE NRTS NETWORK TO\nRADIOTELEMETRY\n32\n4. THE EFFECT OF SAMPLING INTERVAL ON\nTURBULENT ENERGY SPECTRA\n33\n5. SITE EVALUATION PROGRAM\n34\n1.5 Introduction\n34\n5.2 Design Basis Accident Model\n38\n5.3 Average Annual Model\n39\n6. DIRECTIONAL WIND SHEAR UNDER LOW WIND\nSPEED CONDITIONS\n43\n7. APPLICABILITY OF GEOSTROPHIC WINDS TO\nMESOSCALE TRANSPORT\n44\n8. AN APPLICATION OF THE MESOSCALE WINDFIELD\nMETHOD\n50\n9. LONG-TERM TRAJECTORIES DETERMINED FROM\nSINGLE- AND MULTI-STATION MODELS\n52\n10. LONG DISTANCE TRANSPORT AND DIFFUSION\nTO 100 KM\n63\n11. DIAGNOSTIC APPLICATIONS OF WIND SPEED\nAND COMPONENT SPECTRA\n71\niii","Page\n12. FORECASTING AND WEATHER WARNING SERVICE\n80\n82\n13. ACKNOWLEDGMENTS\n83\n14. REFERENCES\nLIST OF FIGURES\nFigure\nMap showing the radar, tower, the partitioning of the\n1.\n4\ncity.\n2.\nSequential tetroon trajectories.\n5\nMean tetroon trajectories across the city determined\n3.\nby averaging five flights each period.\n6\n4.\nTetroon trajectories across the city.\n7\nMean tetroon height (indicated by dots at 3-min inter-\n5.\nvals) upwind and downwind of city center based on five\nflights each period.\n8\nTetroon-derived longitudinal Reynolds stress as a\n6.\nfunction of time for flights over the city (solid line)\nand over rural areas (dashed line).\n9\nThe orothogonal propeller anemometer used in the study.\n12\n7.\nThe orothogonal propeller anemometer in the extended\n8.\n13\nposition.\nWind station network configuration for Oklahoma study.\n15\n9.\nWindfields as a front passed from the Oklahoma meso-\n10.\n17\nnetwork.\nWindfields showing almost a complete rotation.\n18\n11.\nTrajectory plots at 1-hr intervals showing the paths\n12.\nof hypothetical particles advected with 20-min aver-\naged windfield data.\n19\nWindfield (WF) - single station (SS) comparisons.\n20\n13.\nTetroon (lower case) and windfield trajectory compari-\n14.\nsons at 15-min intervals.\n23\nTemperature profile plots from the WKY tower in Okla-\n15.\n26\nhoma City.\nRelative size and location of mesoscale networks in\n16.\nwhich transport has been studied from using network\nwinds in conjunction with tetroon flights.\n28\n17.\nWind station network configuration for Colorado study.\nThe contour elevation shown is in feet.\n29\niv","Figure\nPage\n18.\nThree hourly windfield plots from Colorado mesonetwork.\nGrid spacing is 6.5 km.\n30\n19.\nTrajectory plots from Colorado data. See fig. 10 for\ndetails.\n31\n20.\nu and S spectra sampling interval = 0.58 sec.\n35\n21.\nV and W spectra sampling interval = 0.58 sec.\n35\n22.\nu and S spectra sampling interval = 1.16 sec.\n36\n23\nV and W spectra sampling interval = 1.16 sec.\n36\n24.\nu and S spectra sampling interval = 2.32 sec.\n37\n25.\nV and W spectra sampling interval = 2.32 sec.\n37\n26.\nDesign basis accident plot of X/Q versus cumulative\nprobability for the Forked River Nuclear Power Station.\n40\n27.\nExamples of annual average plots of X/Q versus dis-\ntance for each directional sector for the Forked River\nsite.\n41\n28.\nContour map of annual average X/Q for the Forked River\nsite.\n42\n29.\nAerial photograph showing smoke being released from\nthe surface, 25 m and 61 m on the Grid III tower.\n43\n30.\nDirectional wind shear in time.\n45\n31.\nComputational grid for geostrophic winds in Oklahoma.\n46\n32.\nDifference of computed geostrophic wind from observed\nwind at Oklahoma City.\n49\n33.\nSeries of four trajectories beginning from CPP at\ndates and times shown.\n51\n34.\nComposite 500-mb chart of 7 days resulting in favor-\nable SW winds for testing on the long range diffusion\ntest grids.\n53\n35.\nSurface chart analysis for 1200 GMT, Sept. 18, 1969,\nand associated set of four trajectories released from\nCPP starting at 1300 on the same day.\n54\n36.\nSame as fig. 35, except time and date is 1300 on\nSept. 13, 1968.\n55\n37.\nA grid square intercepting parts of two direction\nrays of a wind rose.\n56\n38.\nThe damping curve for a = 1/8 in the operator\nIf = $0 + aV2P0.\n57\n39.\nTotal hits analysis for Oklahoma City.\n58\nV","Figure\nPage\n40.\nSame as fig. 30, except for release from LOFT at\nthe NRTS.\n59\n41.\nSame as fig. 30, except for release from 'point F'\nat Los Angeles.\n60\n42.\nSeries of windfield plots showing formation of\nlarge eddy, 2000 through 2300 MST, Feb. 9, 1969.\n62\n43.\nField measurement area with wind tower locations\n(dots) and tracer sampling arcs (A,B,C, and D).\n64\n44.\nMethyl iodide measurements and NRTS climatological\ncurves of relative axial concentrations versus down-\nwind distance.\n66\n45.\nLateral dispersion values from NRTS climatology,\nmethyl iodide measurements, and approximations from\nenvelopes of tetroon trajectories.\n66\n46.\nPaired horizontal trajectories at 15-min intervals\nfor tetroons (lower case letters) and windfield ad-\nvected (upper case letters) hypothetical particles.\n67\n47.\nComposite plot of tetroon trajectories No. 3, 4, and 5.\n68\n48.\nComposite plot of windfield-derived trajectories.\n69\n49.\nVertical dispersion values from the NRTS climatology,\nand the effective values determined from methyl iodide\nmeasurements.\n70\n50.\nPlots of time variation during February 1969 of wind\ncomponents and wind speed for the 76-m level of the\ntower at the Central Facilities Area (CFA) of the NRTS.\n72\n51.\nSame as fig. 50, except for 6 m level.\n73\n52.\nSame as fig. 50, for July 1969.\n73\n53.\nSame as fig. 51, for July 1969.\n74\n54.\nRelief map of the Upper Snake River Plain in SE Idaho.\n75\n55.\nRelative locations of the Arnold Engineering Develop-\nment Center (AEDC) tower, relatively flat ground and\nmountainous terrain.\n76\n56.\nEnergy spectra for wind speed and combined spectra of\nthe u and V components of the wind at 61 m on the AEDC\ntower for 8192 hours beginning December 1, 1964.\n77\n57.\nSame as fig. 56, except at 10 m.\n77\n58.\nEnergy spectra for the u and V components with V being\nN-S component.\n79\n59.\nEnergy spectra for the u and V components with V being\nthe NE-SW component.\n79\nvi","LIST OF TABLES\nTable\nPage\n1.\nComparison of Area Covered by Windfield and\nSingle Station Trajectories\n21\n2.\nStatistics on Separation of Particles Advected\nby the Windfield and Particle Movement by the\nSource Wind Only\n22\n3.\nAverage Direction of Tetroon Trajectory with\nRespect to Windfield Trajectory\n24\n4.\nComparison of Tower Winds and Tetroon-Derived\nWinds\n25\n5.\nThe Stations Observations that the Grid Point\nValues of D, z, , and S* were Interpolated From\n47\n6.\nObserved Winds on the Hour at Three Levels of\nthe WKY Tower and the Computed Geostrophic Wind\n48\n7.\nHourly Average Winds and Peak Gusts Recorded\nDuring the Hour Ending at the Time Indicated\n(direction-speed-peak gust)\n81\n8.\nAltimeter Setting Differences in Inches of\nMercury\n82\nvii","","PREFACE\nIn accordance with the letter of agreement of July 13,\n1971, with the U.S. Atomic Energy Commission, Division of\nReactor Development and Technology, the Air Resources Labora-\ntories have continued their study of atmospheric transport\nand diffusion in the planetary boundary layer, micrometeor-\nology, diffusion climatology, and the application of this\nwork to the disposal of radioactive waste gases into the\natmosphere. The research is technically administered and\nsupervised through the Air Resources Environmental Laboratory\nof the Air Resources Laboratories. The work is performed at\nthe Air Resources Laboratories Headquarters in Silver Spring,\nMaryland, and at the Air Resources Idaho Falls Laboratory,\nNational Reactor Testing Station, Idaho. Any inquiry on the\nresearch being performed should be directed to the editor,\nIsaac Van der Hoven, Chief, Air Resources Environmental\nLaboratory, Air Resources Laboratories, National Oceanic and\nAtmospheric Administration, 8060 - 13th Street, Silver Spring,\nMaryland 20910.\nix","ATMOSPHERIC TRANSPORT AND DIFFUSION IN THE\nPLANETARY BOUNDARY LAYER\nAIR RESOURCES LABORATORIES ANNUAL RESEARCH PROGRAM\nREVIEW FOR THE ENVIRONMENTAL SAFETY BRANCH\nDIVISION OF REACTOR DEVELOPMENT AND TECHNOLOGY\nU. S. ATOMIC ENERGY COMMISSION\n1. URBAN PLANETARY BOUNDARY LAYER STUDIES (OKLAHOMA CITY)\n1.1 Introduction\nIn a continuing effort to learn something about the effect that a\ncity has on the low-level windfield, we carried out an extensive obser-\nvational program near Oklahoma City from mid-September to mid-October\n1971. A number of factors led to the choice of Oklahoma City for this\nprogram. The most important ones were a meteorologically instrumented\n500 m tower (the WKY TV tower, with meteorological equipment installed\nand operated by NOAA's National Severe Storms Laboratory); the relative\nflatness of the surrounding terrain so that the city itself could reason-\nably be assumed to be the only disrupting element in the flow; and the\nexpected prevalence of steady, moderate to strong southerly winds, which\ngreatly simplified the design of the observational network and permitted\naccumulation of an \"ensemble\" of observations under similar meteorological\nconditions.\nThe measurement program consisted of four separate parts: (1) fast-\nresponse sensors were mounted on the WKY tower at two levels to accurately\nmeasure the stress tensor and other turbulence statistics needed for\nproper interpretation of the other data; (2) tetroons were released at\nsites upwind from the city, the release points being varied as circum-\nstances dictated, but were always chosen so that the tetroons would pass\nclose to the WKY tower, with the tracking radar situated about 8 km from\nthe tower; (3) pilot balloon soundings (and some radiosondes) were taken\nby a mobile network of five stations arranged in three different configura-\ntions, those of greatest interest being a line along the mean wind direc-\ntion, passing through the heart of the city, and including a station ad-\njacent to the WKY tower; (4) slow-response wind measurements were made on\na continuous basis at five levels of the WKY tower and at about 30 m\nabove the ground with a network of nine stations occupying a region some\n80 X 130 km centered around Oklahoma City.\nThese four measurement activities contributed to six distinct goals:\n(1) to repeat the comparison of pibal and tower wind measurements carried\nout in 1970 (see NOAA Tech. Memos. ERL ARL-28, 1971 and ERL ARL-32, 1972)\nbut with better quality control and greater precision; (2) to determine","the spatial variations in the windfield in the lowest 2 km, as induced by\nthe presence of the city; (3) to compare features of the windfield deter-\nmined from a line of pibal stations with the same features determined\nfrom a series of tetroon flights passing close to that line of stations;\n(4) to compare the \"Eulerian\" covariance of horizontal and vertical velo-\ncity measured with fast-response sensors mounted on a tower with the\n\"Lagrangian\" covariance of horizontal and vertical velocity determined\nfrom a series of tetroon trajectories; (5) to determine the feasibility\nof using standard pibal and radiosonde techniques to determine the verti-\ncal variation of wind, geostrophic wind, and acceleration in the lowest\n2 km; and to test the technique of determining the shearing stress by\nvertical integration of the geostrophic departure; (6) to test a techni-\nque developed at the National Reactor Testing Station (NRTS) for deter-\nmining trajectories from a network of near-surface wind measurements.\nSince analysis of the data is currently underway, it is now most appro-\npriate to describe the program in detail by breaking it down according to\nmeasurement program.\n1.2 The WKY Tower Measurements\nSlow-response measurements of wind speed and direction were made at\nfive levels on the tower - approximately 26, 44, 90, 266, and 444 m above\nthe surface. Slow-response temperature measurements were made at these\nsame levels and also at the surface and 177 and 355 m. These data were\nrecorded continuously throughout the program and were subjected to vastly\nimproved quality control procedures than in 1970. In addition, fast-\nresponse measurements of three wind components, using orthogonal arrays\nof propeller anemometers were made at 177 and 355 m; these were recorded\nin analog form while tetroon and/or pibal operations were in progress.\nThey are discussed in greater detail in section 1.5.\n1.3 Three-Dimensional Tetroon Trajectories\nConstant volume balloon (tetroon) flights near Oklahoma City during\nSeptember-October 1971 provide information on the influence of an iso-\nlated urban area on the horizontal and vertical airflow in unstable and\nrelatively stable conditions. The tetroons were tracked by an M-33 radar.\nLaunch mobility was provided by a large truck fitted out as an inflation\nvan. Tetroon positions were obtained at 1-sec intervals and stored on\nmagnetic tape. These 1-sec readings were later averaged to yield 30-sec\naverage tetroon positions and derived velocities. As usual, transponders\nwere attached to the tetroons to permit accurate positioning at very low\nelevation angles. The tetroons were inflated to float at a mean height\n300 m above ground, since one purpose of the experiment was to compare\nReynolds stresses derived from the tetroons with the stresses derived\nin the conventional fashion from fixed-point instruments (Gill props)\non the 450-m WKY tower.\n2","The solid lines in figure 1 crudely estimate the partitioning of\nOklahoma City according to building type. The small area labeled 1 repre-\nsents the downtown area with buildings as high as 40 stories (120 m),\narea 2 the semibuilt-up industrial and commercial areas with buildings\nabout 10 stories (30 m), area 3 the residential area, and area 4 the rural\narea. The dashed lines indicate terrain height about sea level at 40-m\nintervals. The slight variation in terrain height is mainly associated\nwith the three river systems in the area: the Cimarron River to the north,\nthe North Canadian River that passes through the southern edge of Oklahoma\nCity, and the Canadian River that passes about 30 km south of Oklahoma\nCity, or near Norman, Oklahoma. The locations of the instrumented TV\ntower and the tetroon-tracking radar are also indicated in figure 1.\nFigure 2 shows sequential tetroon trajectories for the first 2 days\nof October. To make the city's influence on the trajectories more ob-\nvious, we have doubled the east-west scale with respect to the north-south\nscale. The dots indicate tetroon positions at 3-min intervals, the flight\nnumbers are given at the ends of the trajectories, and the time (CDT) when\nthe tetroons passed over the center of the city are shown in parentheses.\nAt this location, the wind consistently backed during the day; therefore,\ntetroons released in the morning tended to pass to the east of downtown\nand those released in the evening, west of downtown.\nQuite apparent on morning flights 19, 20, and 24 is the tendency for\nthe tetroons to turn toward the left looking downwind (toward lower pres-\nsure) as they pass over the city; in most cases this is followed by a re-\nturn nearly to the original track. Complete compensation is prevented by\nthe diurnal backing of the wind. In the afternoon, this turning is not so\nobvious, and instead wave-like oscillations occur over the city. The single\nevening trajectory (flight 23) in figure 2 is nearly straight.\nTo obtain a generalized picture of the turning, examine the mean\ntetroon trajectories in figure 3; these are based on five flights passing\nover the city from 0900 to 1200, 1200 to 1500, and 1500 to 1800 CDT. The\nwind speeds are so nearly the same (13 to 14 m sec-1 along all the tracks\nthat the averaging procedure causes little smoothing, but to ensure a mini-\nmum of distortion, the coordinate axis for the averaging was centered on\nthe city's center and not on the launch site. Between 0900 and 1200 CDT, the\ntrajectory turning (backing) over the city at about 300 m amounted to 9°,\nprobably illustrating the additional frictional effect induced by the city.\nBecause of this turning the trajectory beyond the city shifted about 2 km\nlaterally from that shown by the light dashed lines. Inasmuch as trajec-\ntories either side of the city presumably do not undergo such a shift,\ndownwind of the city there is lateral trajectory convergence to the left\nand lateral trajectory divergence to the right of the city looking down-\nwind. This lateral mass convergence and divergence could be compensated\nfor either by ascending and descending motion, respectively, or by an\nincrease in wind speed while it moves downwind to the right of the city.\nWithout simultaneous tetroon flights, it is difficult to tell which is the\n3","320\n360\n360\n360\n400\n440\nTOWER\nNorth\nRADAR\n2\n/\n440\n400\n3\n4\n400\n400\nLAUNCH\n440\n400\n320\nN\n10 KM\n360\nFigure 1. Map showing the radar, tower, the partitioning\nof the city according to building type (see text), and\nthe terrain height in meters MSL (dashed lines).\n4","OCTOBER 2\nOCTOBER I\n26\n(1300)\n23\n(1840)\n22\n#27\n(1550)\n(1510)\n25\n(1110)\n19\n(0920)\nN\n24\n21\n(0910)\n(1330)\n20\n(1120)\nTOWER\nTOWER\nRADAR\nRADAR\nO\n10 km\n10 km\nFigure 2. Sequential tetroon trajectories. The dots are tetroon\npositions at 3-min intervals; flight numbers and passage (LT)\nover city center are indicated at trajectory end. The E-W\nscale is double the N-S scale.\ncase. Between 1200 and 1500 CDT, the mean trajectory backing over the\ncity amounted to only 3° , but the backing began farther upwind and ex-\ntended farther downwind from the city than at the earlier times. Thus,\nin this case also the lateral trajectory displacement amounted to about\n2 km. Note that in early afternoon, even in the mean, wave-shaped tra-\njectory oscillations are set up by the city and appear to extend at least\n50 km downwind. Between 1500 and 1800 CDT, the mean trajectory passing\nwest of downtown turns eastward at a 5° angle downwind of the city, but\nthere is no apparent upstream effect; however, this may be because the\nlaunch site was too close to the city. The impression that late in the\nday the city acts as an obstacle to the flow is confirmed in figure 4 by\nthe mean trajectories for 1800 to 2100 CDT. The two outer trajectories\nare each the average of two trajectories, and under the relatively stable\n5","0900-1200\n1200-1500\n1500-1800\nN\nTOWER\nTOWER\nTOWER\nO\nRADAR\nRADAR\nRADAR\n10km\n10 km\nFigure 3. Mean tetroon trajectories across the city determined by\naveraging five flights each period. The dashed lines illustrate\nthe mean lateral trajectory displacements caused by the city;\nthe mean angular turning is also shown. The E-W scale is double\nthe N-S scale.\nconditions prevailing at this time, note that the air tends to bend around\nthe city somewhat like the flow of water around a rock. The mean angular\nturnings are small, only about 4°, but an opposing lateral trajectory dis-\nplacement of about 2 km occurred each side of the city, resulting in an over-\nall lateral trajectory convergence in the lee of the city of about 4 km.\nOnce again, simultaneous flights would be required to see if this lateral\nconvergence was compensated for by an upward air motion or increase in\nwind speed.\nThe two individual trajectories in figure 4 crossed the center of the\ncity at 1820 and 1950 CDT, respectively. The point of interest is that the\ncity center induced a fairly pronounced wave-shaped oscillation at 1820,\nbut there is no evidence of such an oscillation at 1950. Apparently, as\n6","1800-2100\n28\n(1820)\nN\n.7\n(1950)\nFigure 4. Tetroon trajectories\nacross the city. The trajec-\ntories farthest west and east\nare each based on the average\nTOWER\nof two individual trajectories.\nRADAR\nDots show tetroon positions at\n3-min intervals.\n10km\n10 km\nthe atmospheric stability increases, the effect the city has on the airflow\nat heights of 300 m becomes very small. This is consistent with the exper-\nience at Columbus, Ohio (Angell et al., 1971), where the influence of the\ncity is mostly evident below 200 m.\nFigure 5 shows the mean tetroon height traces for the five flights in\neach of four time periods, whose mean horizontal trajectories appear\nin\nfigures 3 and 4. The heights have been averaged with respect to distance\nfrom the city's center. The city-induced upward air motion is a maximum\nin the late afternoon, but is also apparent during the morning, with\nrespective mean upward motions of 0.7 and 0.3 m sec-1 The point of maxi-\nmum tetroon height moves farther downwind as the day progresses, but in\ngeneral it is centered almost over the city. Because the restoring force\nacts to return the tetroon to its equilibrium (isopycnic) float surface,\nthis result may be misleading. A computer program estimates the air par-\ncel vertical motion from the tetroon vertical motion. This program will\nbe run for all the Oklahoma City flights, but this has not yet been accom-\nplished. Preliminary results indicate, as expected, that the actual air\nparcel vertical motion above the city is considerably larger than that in-\ndicated by the tetroons, and that in general the crest of the air parcel\ntrajectory is several kilometers downwind of the city.\n7","I.O\nW 0.3 ms\nV=13ms\n0.8\n0.6\n0.4\nW 0.4 ms\nV=13ms\n1.0\nFigure 5. Mean tetroon height\n0.8\nupwind and downwind of city\n0.6\ncenter based on five flights\neach period (indicated by dots\n0.4\nat 3-min intervals). Schematic\nrepresentations of the tower and\nOklahoma City buildings are shown.\n1.2\nW= 0.7 ms-\nV=14ms\nMean tetroon speed is shown at the\nright, mean vertical velocity at\n1.0\ncenter.\n0.8\n0.6\n0.4\nV=14ms-\n0.8\n06\n0.4\n40\n-30 South 20\n-10\no\n10\n20 North 30\n40\nALONGWIND DISTANCE FROM CITY CENTER (km)\nAn interesting feature of figure 5 is that the TV tower appears to be very\nnear the trough in the standing-wave - pattern over the city, so that there\nshould be relatively little mean vertical motion at the tower. In other\nwords, by pure coincidence, for conditions prevailing during the experiment\nthe tower was approximately one-half the standing-wavelength from the cen-\nter of the city.\nIn the evening, the tetroons descended slightly over the city, the oppo-\nsite of their daytime behavior. This may be because so little vertical air\nmotion occurred in the evening that the tetroons basically followed their\nisopycnic float surface, which descended over the relatively warm city\n(heat island effect). During the day, the isopycnic surfaces may also dip\nover the city, but this effect on the tetroon height traces is completely\nmasked by the strong vertical air motions occurring in daytime.\n8","One purpose of the experiment was to compare the Reynolds stress mea-\nsured from sequential horizontal and vertical velocities along the tetroon\ntrajectories with the stress measured in the conventional manner on the\nTV tower. The tower data are not yet available so that comparisons are not\nnow possible. However, it is of interest to see how the tetroon-derived\nstress over the city compares with that over the countryside. For this\npreliminary investigation, we assume that the first hour of tetroon flight\nin the southerly flow was basically over the city (after the first hour,\nthe tetroons were near the tower) and that the second hour of flight was\nover the country.\nFigure 6 shows the mean urban and rural tetroon-derived stress as a\nfunction of time. The urban stress is always greater than the rural stress,\nas would probably be anticipated. Furthermore, although the rural stress\nis a maximum in midafternoon, as might be expected because of the strong\nconvection occurring then, the urban stress is a maximum (nearly 2.5 dynes\ncm- 2 ) in the morning. This shift in peak value to the morning is probably\nbecause the mean tetroon trajectory was directly over the downtown area\n(fig. 3). Thus, this implies that the city, particularly its center, in-\nduces an anomalously large downward transport of momentum.\nAn interesting point still to be investigated is the extent to which\nthe standing-wave pattern associated with the city causes the relatively\nlarge downward transport of momentum over the city. This will be estimated\nfrom the cospectra of longitudinal and vertical velocity. This is of\nspecial interest because the momentum flux associated with the standing\nwave cannot show up in the momentum flux obtained at a fixed point, namely\non the TV tower. Thus, the possibility exists that conventional tower esti-\nmates of momentum flux are seriously wrong in locations near cities or hills\nwhere standing-wave patterns are well established. Of course, aircraft\nmeasurements would yield representative flux estimates under such conditions.\n2.5\nURBAN\n20\nFigure 6. Tetroon-derived\nlongitudinal Reynolds stress\nas a function of time for\n1.5\nflights over the city (solid\nline) and over rural areas\n(dashed line).\n10\nRURAL\n0.5\no\n18\n21\n09\n12\n15\nLOCAL TIME\n9","1.4 Pibal - Radiosonde Measurements\nThis program consisted of three different phases. In Phase I, a line\nof five stations was set up, oriented with the prevailing wind from roughly\n25 km upwind of the downtown area to 25 km downwind, passing through the\ndowntown area and past the WKY tower. In Phase II, the five stations were\nsituated in the four corners and at the center of a rectangle roughly 80\nby 120 km, with the center station at Tinker Air Force Base just outside\nOklahoma City. In this phase, radiosondes as well as pibals were taken.\nIn Phase III, five theodolites were set up adjacent to the WKY tower;\neach balloon was tracked by all five theodolites, permitting evaluation\nof each sounding along 10 baselines of varying length and orientation,\nassessment of observer accuracy, and comparison of pibal-determined winds\nwith those measured on the tower. In all three phases, we used only\ndouble-theodolite tracking. Baselines were, wherever possible, at least\n750 m long, and the basic data period was 2 hr during which 20 balloons\nwere tracked at each station. Each balloon was tracked for 10 min and\nreached a height of roughly 1.5 to 2.0 km. To permit tracking at this\npace, two teams of observers were stationed at each site; the first team\ntracked the first balloon, and halfway through this sounding, the second\nteam began tracking the second balloon, etc.\nPhase I was from September 27 to October 1, the synoptic pattern was\nsteady, and the wind blew directly along the line of stations. Data were\ncollected during 11 different 2-hr periods on these 5 days, and because\nof the unchanging overall picture, these 11 periods may be combined into\na single ensemble. Preliminary examination of the soundings indicates a\nfar greater diurnal variation than had been expected, undoubtedly due in\npart to diurnal changes in geostrophic wind associated with the large-\nscale terrain slope (as described, for example, by Bonner and Paegle,\n1970), but also very likely due in part to the thermal and roughness\nfield associated with the urban area.\nPhase III was on October 14 and 15. The weather was good, and data\nwere collected during five 2-hr periods. Examination of these data has\nnot started.\n1.5 Fast-Response Wind and Temperature Fluctuation Measurements\n1.5.1 Objective\nThe principal objective of this program was to estimate the momentum\nand heat flux, using the eddy correlation method, at sufficient height on\nthe WKY tower to represent flow conditions above the surface layer in the\nplanetary boundary layer. The estimates of the momentum flux in particu-\nlar are to be compared with similar measurements derived from balloon\nand profile data gathered near the tower and from tetroon measurements.\nConsiderable emphasis will be placed on the low-frequency fluctuation\nto find long-period oscillations in the measurements of the vertical\n10","component of the wind, which may be induced by the dynamics of flow over\nthe city.\n1.5.2 Instrumentation\nThe selection of a sensor to measure the orthogonal components of the\nwind was determined by the following factors. First, measurements of the\nmomentum flux were desired above 300 m; therefore, a fairly rugged sys-\ntem, requiring only periodic maintenance and calibration was needed.\nSecond, nuniformity of the local terrain and vegetation near the tower\nwould make any measurements below 40 m suspect. In this height range,\nthe long wavelength end of the inertial subrange is about 1/10 the height.\nThus to measure the high frequency contribution to the momentum flux, we\nneeded a sensor with a distance constant of less than 2 m. What is commonly\nreferred to as a Gill propeller-type anemometer, described by Holmes et al.\n(1964), meets these general requirements. Wind tunnel tests reported by\nCamp et al. (1970) show the distance constant is dependent upon the angle\nof attack with a maximum value of 1.2 m at longitudinal directions that\ndecreases to about 0.4 m at transverse flow directions.\nThree individual propeller units (Young, Model Number 27100) were as-\nsembled to form an orthogonal array by constructing a central mounting\nblock (fig. 7). The aluminum block was 10 cm in diameter and 5 cm thick.\nThe arm support holes were precision drilled to guarantee the angle be-\ntween the respective propeller shafts of 90° + 0.1°, the accuracy suggested\nby Kaimal and Haugen (1969) for accurately determining the momentum flux.\nAt the risk of adding slightly to the distance constant, a shaft extender\nwas used on the vertical pointing propeller to unify the response in trans-\nverse flow conditions. The propeller shaft connects to a small generator\n(53 mV m-1 sec-1), and signal wires transmit the voltages to the base of\nthe tower. Two such arrays were constructed.\nObtaining a level array to the accuracy (+ 0.1°) suggested by Kaimal\nand Haugen (1969) at 300 m above ground presented a much bigger problem.\nFollowing the procedures suggested by Dunbar (private communication) a\nsensitive electronic level monitor was constructed. The sensors (Hamlin,\nModel Number 106), variable resistance type bubble levels (fig. 7), were\nmounted on the aluminum mounting block parallel to the axis of the hori-\nzontal propeller shafts. The position accuracy of the electronic level\nsensors was checked and found to represent the horizontal plane defined\nby the propeller shafts within the limits specified above. An a-c bridge\ncircuit was constructed to measure the resistance of the level indicators\nThe bridge circuit was \"portable,\" and the level indicators were useful\nin positioning the array after calibrations. Sufficient signal wires were\navailable to allow monitoring of the level sensors from the base of the\ntower at the beginning and end of each run.\nA sensitive thermometer, consisting of a five-junction thermopile\n(No. 40 copper/constantan), was constructed to measure temperature fluc-\ntuations to allow us to estimate the heat flux. Both the sensing junction\n11","Figure 7. . The orothogonal propeller anemometer used in the study.\nThe shaft extender is clearly visible atop the vertical pointing\nanemometer and the bubble levels are to the right at the base of\nthat component.\nand the reference junction were mounted in an aspirated radiation shield\n(Climat, Model Number 061-1) similar to that used to shield the tempera-\nture gradient measuring system. The reference junction was enclosed in\nan epoxy mass to lengthen its time constant to agree with that of the\ngradient sensor. Tests showed that when mounted in the aspirator, the\nsensor junction had a time constant of about 1 sec (1/e time), and the\nreference junction was about 70 sec. Airflow in the shield was computed\nto be 3 m sec-1. .\n1.5.3 Exposure\nAs discussed in a preceding section, one of the important objectives\nof this study was to measure the effect of a city on a relatively uniform\n12","flow. Thus, on the WKY tower, located NNE of downtown Oklahoma City, the\nprimary measurements (wind and temperature) were taken from the SW side.\nThe support for the orthogonal wind system was constructed to extend from\nthe tower arms used to mount the profile system, as shown in figure 8.\nIn this position, the center of the array extended 4.2 m from the face of\nthe tower and had its major axis pointing in a 240° + 2° direction. Thus,\nwinds blowing down the center of the orthogonal wind system blow from 195°.\nThroughout the measurement program, wind and temperature fluctuation sen-\nsors were 354 m above ground. The aerovane was removed to avoid any pos-\nsibility of interference. The second system was 158 m above ground through-\nout most of the test; temperature fluctuations were not measured at the\nlower level.\nFigure 8. The orthogonal propeller anemometer in the extended position.\nAt this time the propeller anemometer and the aerovane were being com-\npared, normally the aerovane was removed to avoid interference.\n13","The two primary factors that influence the accuracy of the wind com-\nponent measurements, deviation from a cosine response and flow interfer-\nence caused by the tower, are wind direction dependent. Maximum accuracy\nwas obtained when the wind direction was 195° + 40°, or within 5° of being\ntransverse to one of the horizontal sensors. According to Gill et al.\n(1967), the effect the tower has on the measurements is less than 2 per-\ncent in this direction sector, if we assume the anemometer extends 3 m from\nthe tower. This approximately true since a tower is about 3.1 m on a side\nand the anemometer extends 4.2 m. Tower effects do not exceed 5 percent\nuntil wind direction fluctuations exceed 195° + 80°. Winds from 310°\nthrough north to 80° cannot be analyzed due to tower interference. The\ncalibration error caused by a deviation from a cosine response at trans-\nverse attack angles, when the propeller is changing direction of rotation,\ncan be minimized when a data run can be subdivided into periods when the\nflow is longitudinal and when it is transverse. Such subdivisions were\ngenerally possible in the stronger wind conditions; major problems were\nencountered though when winds were light and variable. Although an analy-\ntic expression for the calibration has been proposed by Drinkrow (1972),\nit was not used; rather, a table of values based on figure 3 from Holmes,\net al. (1964) was used to correct for deviations from a cosine response.\n1.5.4 Signal Conditioning, Recording, and Calibrating\nAll signals were transmitted to ground level by shielded cables;\nnonetheless, a considerable amount of high frequency noise was acquired.\nA simple first-order filter, with cut-off frequency of 13 Hz, limited the\nnoise on all six anemometer signals. The gain of the amplifier and the\ninput bias were changed from one run to the next in order to maintain an\noptimum signal level at the input to the recorder. All signals, includ-\ning a zero level and bias voltage, were recorded in FM mode on a 14-\ntrack analog magnetic tape recorder. The six wind channels were also\nmonitored on a multichannel stripchart recorder.\nA voltage calibration, zero and + 200 mV, was performed on each\namplifier at the beginning and end of each run in order to document gain\nand bias settings. At four times during the tests, a calibration signal\nwas applied from the anemometer using a constant speed (1800 rpm) motor\nto spin the anemometer shafts; propellers were removed during this pro-\ncess. The range in the values from each anemometer was less than 1/2\npercent of the calibration value. Thus a single set of calibration vol-\ntages was used for all of the records.\nThe sensors operated from September 24 through October 12, 1971, and\nrecorded 77 hours of data, the mean wind direction was between 160 and\n220° for 47 hr of this period. Of the remaining 30 hr of data, approxi-\nmately one-half are not useful because directions were more northerly\nand tower influences will not be negligible. All tapes were digitized at\nthe NRTS computer facility. The digitizing rate was 10 observations/sec,\nand the resolution is + 1 part in 8192. The data will be subdivided into\n14","20-min blocks for editing and computing of basic statistics - the mean,\nvariance, skewness, and kurtosis. The spectra and cospectra will be com-\nputed for longer periods, in most cases 60 to 100 min long.\n1.6 Mesoscale Windfield and Transport Analysis\nThe network of wind stations used in this experiment was set up speci-\nfically for this study. Except for the NRTS mesonetwork, it was ARL's first\nattempt to establish a mesoscale network of wind observations for a field\nexperiment. The wind observations obtained for the Los Angeles experiment\n(see NOAA Tech. Memo. ERL ARL-32, 1971) were obtained from stations pre-\nviously established for other purposes; this caused some of the data to\nbe of questionable value because of sensor height and location.\nSince the area surrounding Oklahoma City is moderately populated, it\nwas possible to mount the wind sensors approximately 30 m high on existing\ntowers. Figure 9 shows the positions of the eight wind stations in the\n90 X 130 km area centered on the WKY tower. Wind data were also extracted\nat 44, 266, and 444 m for use in the analyses. The configuration of the\nnetwork provided optimum coverage of south-to-north tetroon flights.\n0\n5\n10 mi.\nScale\nterrain ht. < 1200 ft. ms\n850.\nFigure 9. Wind station network\nconfiguration for Oklahoma\nstudy. Flat topography is in-\ndicated by the 1200-ft contour.\nWKY tower\n1437 .\nterrain ht. > 1200 ft. msl\n15","The local topography provided an interesting contrast to the terrain\ncharacteristics of the upper Snake River Plain. The height difference\nbetween the highest point on the grid (about 440 m) to the lowest point\n(about 260 m) is about 180 m in a distance of about 90 km. This repre-\nsents an average slope of about 2 X 10-3, or about 1/5 of the slope over\nthe NRTS mesonetwork, and would indicate a more gently varying terrain.\nOn a scale of less than a kilometer, however, the Oklahoma terrain may\nvary by over 30 m in a distance of 400 m, which is considerably rougher\nthan the NRTS terrain. The most noticeable terrain feature within the\nboundaries of the grid is the protrusion of the slightly higher ground\nfrom the west. Oklahoma City is on a peninsula of this slightly higher\nground with the terrain sloping downward and away from the city to the\nnorth, east, and south.\nThe wind data set from the mesonetwork was from September 21 through\nOctober 12, 1971. The wind speed and direction on the strip charts were\naveraged over 20-min periods and recorded on punch cards for editing and\ntransfer to disc storage.\nWindfield plots every 3 hr, similar to those for the NRTS network,\nwere produced and examined. There were no wind patterns that could be\nrecognized systematically with mesoscale features observed in the data\nset. Under the conditions of a very weak synoptic-scale pressure gradient,\nthere seemed to be a slight influence of the higher ground to the west, as\nindicated by the nighttime westerly winds. This phenomenon would probably\noccur more during July and August when daytime heating is greater.\nThe most severe horizontal variations in the windfields of this data\nset seem to be associated with the fronts passing through the area. An\nexample of this is shown in figure 10. For 6 days previous to this series,\nthe windfields showed southerly flow with very little horizontal variation.\nShortly after 0400 CDT October 2, a weak front passed the northwest corner\nof the grid and proceeded southeastward causing the spatial variation shown\nin the series of plots. This frontal system was accompanied by a consider-\nable amount of thundershower activity, which would account for much of the\nhorizontal variation. By 0700 October 3, the front was 80 km southeast\nof Oklahoma City and the windfield pattern is once again quite uniform,\nbut with winds from the north.\nThe overall conclusion one might draw from the windfield plots of this\ndata set is that the primary influences on the flow through the grid were\nthe synoptic-scale pressure patterns and frontal systems. Figure 11 shows\nan interesting circulation pattern, of a larger scale than the grid, pass-\ning through the area on October 11. A frontal system oriented east-west\nthrough southern Kansas moved slowly toward the area during the period but\ndid not pass through, while the flow through the area rotated through 360°.\nThe transport of hypothetical particles from the WKY tower for the\n22-day period of the experiment is depicted with the same type of trajec-\ntory plots that were produced for the NRTS (Wendell, 1970). The high\n16","0700 10/02/71\n1000 10/02/71\n1900 10/02/71\n2200 10/02/71\n1300 10/02/71\n1600 10/02/71\n1000 10/03/71\n0100 10/03/71\n0400 10/03/71\n0700 10/03/71\n2200 10/03/71\n1300 10/03/71\n1600 10/03/71\n1900 10/03/71\nFigure 10. Windfields as a front left the Oklahoma mesonetwork.\nGrid spacing for the random-to-grid interpolution is 8 km.\nWind vector scaling is 11 m sec-1 per grid unit.\n17","1000 10/11/71\n2200 10/11/71\n1900 10/11/71\n1600 10/11/71\n1300 10/11/71\n1000 10/12/71\n0400 10/12/71\n100 10/12/71\nFigure 11. windfields showing almost a complete rotation.\nSee figure 10 for other details.\npercentage of southerly flow is indicated by the fact that 18 out of 42\nplots had all 12 trajectories exit the northern boundary A 4-day period\nthat this occurred is shown in figure 12. A close examination of these\nplots reveals a persistent phenomenon. For the daytime releases, begin-\nning at 0800 CDT, the trajectories of the particles released at 1-hr in-\ntervals shift toward the west, but for the night releases the trajectories\nshift toward the east. This behavior was observed for all the 18 cases of\nsustained southerly flow. It would seem to be related to the diurnal os-\ncillation in the southerly flow observed in long-term averages by Bonner\n18","WKY 0800 09/25/69\nWKY 2000 09/25/69\nWKY 0800 09/26/69\nWKY 2000 09/26/69\n1211\n43\n+\nup\nfor\nN\nWKY 0800 09/27/69\nWKY 2000 09/27/69\nWKY 0800 09/ 28/69\nWKY 2000 09/28/69\n121 10 8542\n12\n3\nBill\nUK\nFigure 12. Trajectory plots at 1-hr intervals showing the paths of hypo-\nthetical particles advected with 20-min averaged windfield data. Each\nplot contains 12 particle trajectories with the release time for the\nfirst particle shown at the top of the plot. The arrow at the top of\neach plot indicates the general direction of the wind change shown by\nthe trajectories.\nand Paegle (1970) They attributed the oscillation to the alternate heat-\ning and cooling of the sloped terrain. A reexamination of the windfield\nplots indicates that the phase of the oscillation agrees well with their\nobservations.\nAn oscillation of this type would have important implications in both\naccidental and routine releases of an effluent. For the emergency situa-\ntion, the ability to anticipate a 15 to 30° turning of the wind would be\nimportant. Under a sustained and controlled effluent release, the rela-\ntionship of the oscillation to the diurnal variation in stability would\nbe helpful in determining the area that would receive the highest long-\nterm dose.\n19","of the remaining 24 trajectory plots, three indicated sustained\neasterly flow, seven sustained northerly flow, four sustained westerly\nflow, and 10 a dispersion over a significant portion of the grid. In\nthese 10 sets only five contained evidence of the spreading being caused\nby curved streamlines. This observation is confirmed by the differences\nshown in comparing the windfield trajectories with those produced by the\nsource wind only.\nTrajectory plots for the same data set were produced using the WKY\nwind as if it applied over the whole grid. The results of a comparison\nof the areas covered by the two sets of plots are shown in table 1. Ex-\namples of these comparisons are shown in figure 13. The first two show\npoor agreement in terms of area covered. The third comparison (0800 CDT,\nOctober 11, 1971) agrees well in spite of the large spread in the trajec-\ntories. This release period corresponds to the occurrence of the rotating\nwindfield discussed above (fig. 11). An interesting feature in this\nsituation is that a seemingly significant amount of spatial variation\noccurred caused by curved streamlines from 1000 CDT on the 11th to 0700\nCDT on the 12th, but it is lateral to the direction of transport and\ncauses little discrepancy in the single-station trajectories. The last\nWKY 0800 10/11/71 WF\nWKY 2000 10/11/71 WF\nWKY 2000 10/02/71 WF\nWKY 0800 10/04/71 WF\nWKY 0800 10/11/71 SS\nWKY 2000 10/11/71 SS\nWKY 2000 10/02/71 SS\nWKY 0800 10/04/71 SS\n12th\nFigure 13. Windfield (WF) - single station (SS)\ncomparisons. See fig. 12 for other details.\n20","Table 1. Comparison of Area Covered by Windfield\nand Single Station Trajectories\nNumber of\nPercent of\nArea Agreement\nCases\nTotal\nGood (90% to 95% mutually covered)\n27\n64\nFair (70% mutually covered)\n9\n22\nPoor (<70% mutually covered)\n6\n14\ncomparison in figure 13 is a continuation of the October 11 release. If\nwe consider area coverage alone, there is fair agreement in the plots, but\nindividually, the last nine of the 12 trajectories show significant\ndifferences.\nTo compare trajectories of individual particles computed with wind-\nfield and single-station data, we calculated separation distances and\nstored them according to distance traveled by the particle in the wind-\nfield. Some statistical results of these comparisons are shown in table 2.\nThe number of particles on the grid is strongly dependent on the grid geom-\netry. The distances from the source to the boundaries are about 43 km east\nand west, 65 km to the north and south, and about 77 km to the grid corners.\nThere is a distinctive drop in the number of particles available for com-\nparison as each of these distances is exceeded, indicating a high percen-\ntage of relatively straight trajectories. The separation statistics of\ntable 2 indicate that, for the period of this data sample, inside a 32-km\nradius a particle being advected by the source wind would be within 8 km\nof its position in the windfield for at least 70 to 80 percent of the time.\nAt a distance of 48 km, with about 89 percent of the particles still on the\ngrid, the average separation is 14 km and growing rapidly. This would\ncause serious problems only for unexpected short releases of material.\nThe favorable area comparisons would show single-station transport to be\na reasonable approximation for long-term controlled releases.\nEven though it was not their primary purpose, the tetroons flown\nthrough the grid provided a valuable source of information on the flow\nabove the tower network. There were 54 flights that were compared with\nsimultaneous windfield trajectories. The average heights of these\nflights varied as follows: 24 flights at or below 300 m (above the ground),\n25 flights above 300 m, and five flights above 610 m. The average height\nfor all the flights was 380 m above the ground. A trajectory plot was\nproduced for each tetroon flight including also the trajectory of a hypo-\nthetical particle carried by the windfield. Information was printed at\n15-min intervals on the positions, separation, speed, and direction of the\ntetroon and particle. Figure 14 shows examples of these plots. The primary\n2F","Table 2. Statistics on Separation of Particles Advected\nby the windfield and Particle Movement by the Source\nWind Only.\nPercent of Particles in Four\n*D\nN\nS\nS min\nS\no S\nWF\nP\nmean\nmax\nSeparation Categories\n(km)\n(km)\n(km)\n(km)\n(km)\n8\n8-24\n24-40\n40 km\n0800 to 1900 CDT\n16\n126\n1.6\n2.6\n21.6\n0.0\n96.8\n3.2\n0.0\n0.0\n32\n126\n7.4\n10.6\n66.4\n0.5\n78.6\n14.3\n4.0\n3.2\n48\n112\n14.8\n18.7\n98.0\n1.8\n41.1\n50.0\n1.8\n7.1\n64\n75\n22.7\n23.5\n121.0\n3.7\n13.3\n58.7\n20.0\n8.0\n80\n17\n44.5\n30.2\n117.0\n9.7\n0.0\n35.3\n17.6\n47.1\n96\n13\n60.8\n28.4\n114.0\n17.4\n0.0\n7.7\n23.1\n69.2\n2000 to 0700 CDT\n16\n120\n1.4\n1.3\n5.7\n0.2\n100.0\n0.0\n0.0\n0.0\n32\n120\n8.0\n9.9\n61.2\n0.2\n68.3\n26.7\n2.5\n2.5\n48\n111\n14.0\n15.3\n91.9\n0.3\n47.7\n35.1\n2.6\n4.5\n64\n53\n28.6\n20.8\n78.6\n1.6\n22.6\n24.5\n26.4\n26.4\n80\n8\n49.1\n19.0\n90.0\n21.3\n0.0\n13.5\n35.0\n62.5\n96\n6\n37.4\n16.2\n57.4\n17.9\n0.0\n33.3\n16.7\n50.0\n*\nD\n- distance traveled by windfield advected particles.\nWF\nNp\n- number of windfield particles still on grid.\nS\n- particle separation.\ntransport factors are wind speed and direction shear with height. Since\nthe tetroons were generally well above the level of the network sensors, the\naverage speed of the tetroons was always greater than the average speed of\nthe particle. A logarithmic extrapolation of the average speed of the\nparticle in the windfield to the average tetroon height was within + 15\n-\npercent of the average tetroon velocity in 47 percent of the cases. The\nerrors in the extrapolation seemed to be fairly evenly distributed on the\nhigh and low side, providing no reason to bias the extrapolation for better\naccuracy. Thus, this series of primarily daytime tetroon flights indicates\nthat a logarithmic extrapolation of the 30-m wind speed in the network to\nthe height of interest should generally provide a reasonable estimate of\nthe transport speed in the lowest 610 m.\n22","F26 1219 10/01/71\nF24 0835 10/01/71\nF25 1026 10/01/71\nF27 1437 10/01/71\nF28 1744 10/01/71\nF29 1140 10/02/71\nFigure 14. Tetroon (lower case) and windfield trajectory\ncomparisons at 15-min intervals. Release times shown\nat the top of the plots.\nFor transport, a more important consideration than speed shear is\ndirection shear. Varying amounts of direction shear were observed in the\ntrajectory comparisons. For the 50 applicable cases, the breakdown on\nthe average direction difference between the tetroon trajectory and the\nwindfield generated trajectory is shown in table 3. These results verify,\nin a general sense, the basic physical principle that surface friction\nwill cause cross-isobaria flow in the lower levels when the wind turns to\nthe right with increasing height. In a southerly wind, almost 70 percent\nof the flights show an average of an 8° shift to the right. There seemed\nto be no significant correlation of the amount of shear with average\n23","Table 3. Average Direction of Tetroon Trajectory with\nRespect to Windfield Trajectory.\nWind Direction\nLeft\nCentered\nRight\nTotal\n1* (5)**\n23 (8)\nSoutherly\n11\n35\n2 (10)\n5 (12)\nNortherly\n3\n10\n2 (12)\nWesterly\n2\n0\n4\nEasterly\n1\n1\n5 (10)\n28 (9)\nTotal\n17\n50\n* number of cases\naverage number of desgrees deviation to the left or\nright of the windfield trajectory\ntetroon height, but for the southerly flow the variation around the\naverage was only 10°. The maximum shear encountered in the 50 cases was\n20° These results would indicate that at least during the fall between\nsunrise and sunset the flow in the lowest 600 m can be reasonably well\napproximated by applying a logarithmic extrapolation from the 30 m speed\nand a correction of 10° to the right for heights above about 200 m.\nFactors such as frontal systems, strong inversions, and topographic\nfeatures will complicate the vertical structure of the flow (flight 29\nin fig. 14). There were no tetroon flights at night to compare with wind-\nfield trajectories during inversion conditions; however, it is possible\nto investigate the vertical shear of the horizontal wind by using the\nwind data at the 266 and 444 m levels on the WKY tower.\nTo check the consistency of the wind data from the WKY tower with the\nwind data from the tetroons, we compared the speed and direction of all\ntetroons that passed within a 16-km radius of the tower with tower wind\nspeed and direction interpolated to the tetroon height at its closest\nposition to the tower (table 4). The minor bias toward a negative dif-\nference in the directions should not be considered serious because the\ndirections from the tower data are 20-min average values read to the\nnearest 10°. For some reason, the anemometers on the tower generally\nreport the speed slower than is indicated by the tetroon. The average\ndifference is about 34 percent of the average speed differences shown be-\ntween the tetroon trajectories and the windfield trajectories. This same\ntype of speed difference has been found between the tower winds and those\ndetermined by pilot balloon. An investigation for a satisfactory explana-\ntion of this problem is still underway.\n24","Table 4. Comparison of Tower Winds and Tetroon-Derived Winds.\nThe values shown are obtained by subtracting tetroon data\nfrom tower data. Wind direction is the direction from\nwhich the wind was blowing.\nType of\nNumber of\nMean Direction\nNumber of\nMean Speed\nDifference\nCases\nDifference\nCases\nDifference\n(°)\n(m sec-1)\nPositive\n17\n6\n5\n1.8\nNegative\n26\n-7.4\n35\n-1.8 -\nNone\n0\n0\n3\n0\nTotal\n43\n-2\n43\n-1.2\nIn anticipation of an analysis of the vertical wind shear and the ver-\ntical oscillations exhibited by the tetroons, we developed a method of pre-\nsenting the history of the temperature profile measured on the WKY tower.\nThe 20-min averaged values of these temperature profiles were plotted on\nmicrofilm - three to a graph - with a time interval of 2 hr. The last pro-\nfile on one plot was used as the first profile on the next; this allows\nthe heating or cooling during a 2-hr interval to be examined continuously.\nSix plots were required to cover 1 day, and 2 days can be shown on one page.\nThe history of the temperature profile on the WKY tower was plotted for\nSeptember 27, 1971, through October 12, 1971. Examples are shown in\nfigure 15.\nTwo contrasting histories have been selected for demonstration. The\nfirst series is for a day (September 28) that has been preceded by 2 days of\nsoutherly flow bringing warm moist air from the Gulf of Mexico. Clouds\nduring the night prevented the formation of an inversion by radiative cool-\ning and very little heating was required to bring about neutral stability\nby 1000 CDT. The second series is for a day that was preceded by 2 days\nof northerly flow, behind a passing cold front, bringing cold dry air from\nthe midwest. Here a strong inversion is developing during the early morn-\ning; it is then strengthened by warming above 200 m between 0400 and 0800\nCDT. The stability condition during this time contrasts sharply with that\non September 28. The rapid heating of the lowest 200 m and cooling of the\nupper 300 m between 0800 and 1200 CDT on October 11 suggests turbulent\nheat transfer between these layers in addition to the normal heating for\nthis period. This might be related to the breakup of a low-level jet,\nwhich was observed in the tower winds and is somewhat substantiated by\nsome sporadic vertical oscillations in a tetroon trajectory between 0930\nand 1000 CDT. The oscillations had an amplitude of 100 m and a period of\n25","(a) 09/28/71\nsee\nsee\n500-\n0703\n@@@@\n& 1100\n2028\n0385\n0505\n&\n0703\n2035\n2305\n0105\n0385\n2027\n2040\n2036\n14\n1942\n4001\n400\n400\n300\n300\n300\n2127\n44\n2028\n2025\n1934\n2029\n1937\n200\n200\n200\n4.\n100\n100\n100\n1918\n1917\n2019\n1919\n1819\n44\n&\n0\nI\n.\n0\n10\n15\n20\n25\n30\n0\n25\n30\n20\n30\n25\n5\n10\n15\n20\n5\n10\n15\nTEMP\nC\nTEMP\nC\nTEMP\nC\n500\n500r\n500-\n1900\n2100\n&\n2300\n1500\n1700\nA\n1900\n1500\n1935\n1100\n1300\nA\n1022\n1829\n1922\n1924\n2024\n400\n400\n400\n300\n300\n300\n1932\n1927\n930\n2025\n2025\n2022\n200\n200\n2004\n100\n100\n100\n1919\n1815\n1815\n1919\n1914\n2021\n8+\n25\n30\n@\n5\n10\n15\n20\n15\n20\n25\n30\n$\n10\n26\n30\n10\n19\n20\nTEMP\nC\nTENA\nc\nTEMP\nc\n(b) 10/11/71\n500\nsee\nsee-\n0700\n0900\n1100\n&\n0300\n0500\n& 0700\n2909\n2300\n0100\na 0300\n2725\n2820\n2525\n2627\n2224\n400\n400\n400\n300-\n300\n300\n2910\n2724\n2633\n2436\n2586\n233\n200\n200\n@@@\nde\n100-\n100\n100\n2709\n2217\n2215\n2218\n2218\n2220\nN.\n25\n30\nLOU\n30\n10\n15\n20\n15\n20\n25\nis\n20\n25\n30\n10\n10\nTEMP\nC\nTEMP\nC\nTEMP\nC\n500\n500\n500-\n1900\n2100\nA 2300\n1500\n1700\nA\n1900\n4 1500\n1100\n1300\n1811\n1601\n1706\n2906\n3101\n501\n400\n400\n400\n300\n300\n300\nof\n1413\n1204\n1208\n3201\n/ 60\n3001\n200\n200\n2004\n100\n100\n100\n1114\n3302\n3008\n1201\n0808\n0814\n)\n0\n*\n30\n20\n25\n30\n10\n15\n20\n25\n5\n10\n15\n10\n15\n20\n25\n30\nTEMP\nC\nTEMP\nC\nTEMP\nC\nFigure 15. Temperature profile plots from the WKY tower in Oklahoma\nCity. The identifying symbols and times are at the top of each\ngraph. The slanted dashed lines are the dry adiabatic lapse rate.\nWind direction and speed at three levels are shown for the second\nand third profiles of each plot. The first two digits are wind\ndirection in tens of degrees and the last, wind speed in miles\nper hour.\n26","about 7 min, which is greater than expected for the thermal stability that\nexisted at the time.\nIn summarizing the results of the preliminary analysis of the data\nfrom the Oklahoma experiment, we must emphasize that the 3-week sample\nin the fall would probably not represent conditions in the extremes of\nwinter or summer; however, the sample probably represents about 6 to 7\nmonths of the year. The terrain could also be considered to be fairly\nrepresentative of a large area of the central United States but not of the\nwestern and eastern portions of the country. The terrain appears to have\na much less pronounced influence on the mesoscale flow patterns here than\nis observed at the NRTS. This suggests average transport might be rea-\nsonably well estimated from the source wind only. Trajectory pattern com-\nparisons for both types of advection show the single-station advection does\nfairly well in covering the same area as the windfield advection for a 12-\nhr release. However, comparing individual trajectories shows differences\nbecoming large for short-term unexpected releases. The 20 to 30° diurnal\noscillation is southerly flow that shows up so dramatically in the trajec-\ntory plots is primarily a temporal variation, but an important phenomenon\nof which to be aware in terms of transport. The tetroon and windfield\ntrajectory comparisons indicated that during daylight, a logarithmic ex-\ntrapolation of tower speed and a 10° correction to the right of the 30-m\nwind direction will give a reasonable first guess of the transporting winds\nin the lowest 600 m of the atmosphere.\nFurther and more detailed analysis of the rest of this unique data\nset should provide valuable information in several areas. For example,\nthe available surface synoptic data and upper air soundings should be used\ntogether with the mesonetwork, tetroon, and tower data to determine the\nworth of this routinely available data in the determining of the meso-\nscale transport in moderately populated areas. There can be significant\ndifferences in the wind representation between consecutive 20-min averages\nfrom strategically positioned sensors at 15 to 30 m and 1-min averaged\ndial readings once an hour from a sensor 3 to 6 m above an airport ter-\nminal. The temperature and wind data available on the WKY tower, along\nwith the tetroon and synoptic weather data provide an opportunity to study\ntheir detailed behavior in variance stability conditions. This informa-\ntion should also provide a more accurate estimate of vertical wind shear\nfor practical use.\n2. PRELIMINARY WORK ON HASWELL, COLORADO, MESOSCALE STUDY\nThis experiment is being planned in conjunction with a series of\nexperiments that will be conducted in August 1972 by the Wave Propagation\nLaboratory of ERL. Their experiments involve monitoring several atmos-\npheric properties with sophisticated indirect measurement techniques. Our\npurpose is to collect supplementary and comparative information by direct\ntechniques and also the benefit of the data gathered by the indirect\nmethods. The experiment will also provide an opportunity to study mesoscale\n27","wind patterns and transport in an area having topographic characteristics\nsomewhere between those of the NRTS in Idaho and those of central Oklahoma.\nThe site of the WPL experiment is within a few kilometers radius of their\n154-m tower, about 6 km south of Haswell Colorado. The 80- by 80-km\nmesogrid for this study is centered on the WPL tower and is shown in\nfigure 16 along with the NRTS, Oklahoma, and Los Angeles grids.\nThe mesoscale wind network was set up in late May 1972 to provide\nlonger data sample than would have been available during the 2-week WPL\na\nexperiment in August. Because of the sparse population in southeastern\nColorado, existing towers could not be counted on to provide positions for\nelevated sensors. For this reason, enough tower sections for nine stations,\nalong with the sensors and recorders, were taken to the site of the ex-\nperiment. Techniques developed for efficient erection of 15-m towers and\nexperience in sensor and recorder installation gained in Oklahoma accounted\nfor the rapid deployment of nine wind stations over a 6400-km2 area, plus\n50°\n1300\n1200\n110°\n80°\n90°\n100°\n40.\n30°\nFigure 16. Relative size and location of mesoscale networks in which\ntransport has been studied from using network winds in conjunction\nwith tetroon flights.\n28","a sensor atop the 150-m WPL tower. Site location and installation by a\nfive-man crew required 7 days for the entire network.\nThe size of this network and the configuration of the stations were\ndesigned to provide better support to the WPL program than would have been\nprovided by the type of network used in Oklahoma. The station locations\nand three crude height contours over the grid are shown in figure 17. The\nmean slope between the lowest and highest points on the grid is 4.3 X 10-3 ,\nwhich is about twice that found for the Oklahoma grid and half the value\nfor the NRTS grid. In terms of the height variation over short distances,\nthe Colorado grid is quite smooth compared with the Oklahoma grid. The\nmajor terrain features over the Colorado arid are the general downward\nslope from the northwest resulting in a broad ridge. The southern boundary of\nthe grid lies in the Arkansas River Valley with a moderate amount of trees,\ngrass, and farm crops while the northern boundary lies in dry, almost bar-\nren country with meager vegetation and pasture land.\nThe wind station data, including the 15- and 150-m levels on the WPL\ntower, are recorded in the same format as for the Oklahoma data. The\ndata from the strip charts were again averaged over 20-min intervals. Temp-\nerature profiles on the WPL tower will be available for only the 2-week\nperiod of this experiment. We will fly tetroons in pairs during this per-\niod, and this should provide even more information on the vertical struc-\nture of the flow than the individual flights in Oklahoma.\nSome of the wind data taken in early June have been read, edited and\npunched on cards. A sample of the windfield plots for the data is shown\nin figure 18, and a sample of the trajectory plots is shown in figure 19.\nThree terrain contours have been included with each plot to aid in cor-\nrelating variability in the wind pattern with terrain features. For the\n4800\n4400\nFigure 17. Wind station network\nconfiguration for Colorado\nWPL\nstudy. The contour elevation\nTower\nis shown in feet.\no\n5\n10 mi\nScale\n4000\n29","1000 06 17/72\n0400 06/17/72\n0700 06/17/72\n0100 06/17/72\n1900 06/17/72\n17/72\n06/17/72\n160Q 06/17472\n00/06/18/72\n0700 06/18/72\n1000 06/18/72\n0100 06/18/72\n2200 06/18/72\n1300 06/18/72\n1600 06/18/72\n1900 06/18/72\n0700,06/19/72\n1000 06/19/72\n0100 06/19/72\n06/19/72\nFigure 18. Three hourly windfield plots from Colorado\nmesonetwork. Grid spacing is 6.5 km.\nThe scaling\nfor the wind vectors is 11 m sec-1 for one grid unit.\n30","WPL 0800 06/16/72\nWPL 2000 06/16/72\nWPL 0800 06/17/72\nWPL 2000 06/17/72\nB6\n12\nDC\nWPL 0800 06/18/72\nWPL 2000 06/18/72\nWPL 0800 06/19/72\nWPL 2000 06/19/72\nGF\n12\n0\n$\nWPL 0800 06/20/72\nWPL 2000 06/20/72\nWPL 0800 06/21/72\nWPL 2000 06/22/72\n5\n6\n1\n12\n9\nWPL 0800 06/23/72\nWPL 2000 06/23/72\nWPL 0800 06/24/72\nWPL 2000 06/24/72\nIt\n8\n919#2\nFigure 19. Trajectory plots from Colorado data.\nSee figure 10 for details.\nbrief period shown, , there seems to be more spatial and temporal variability\nin the flow than any period of similar length in the Oklahoma data sample.\nHowever, the time of year is not the same and the entire Oklahoma sample\nwas quite short. As the Colorado data sample extends into the autumn and\nthe supplementary tower and tetroon data become available, a comprehensive\ncomparison should be possible.\n31","3. CONVERSION OF THE NRTS NETWORK TO RADIOTELEMETRY\nThe telemetered wind stations have been strategically located on a\nrectangular grid 220 X 100 km orientated NE to SW in the upper Snake River\nPlain. Seven stations are within the boundary of the NRTS and 17 are off-\nsite installations. Temperature sensors are also to be installed at many\nof these locations.\nEach station within the network is scanned at 6-min intervals for\nwind, radiation, and temperature data. A PDPB/S computer interfaced with\na Kennedy tape recorder records the sequence on magnetic tape. Each 6-min\nsequence composes one physical record of information written on the magnetic\ntape in ASCII (American Standard Code for Information Interchange) code.\nAn on-line teletype can also be activated to monitor the network during\nsequencing.\nSeven-track tapes (200 BPI) that record the telemetered data are\nchanged on a routine basis and processed by the IBM 360-75 computer through\nthe remote input terminal located at Central Facilities Approximately\n40 m of tape is used daily, however, the tapes are changed three times a\nweek as a precautionary measure against equipment failure. The FORTRAN\nprogram for processing the telemetered wind and temperature data has been\ncompiled to machine language and these instructions have been written for\ndisk for direct access to the IBM 360-75.\nContents of the data tape are transmitted via telephone wires from the\nremote input terminal directly to a disk pack at the computer center, using\na special tape READ routine. This routine (MR61) is accessed by Job Control\nLanguage (JCL). Upon execution of the job step, the telemetered data re-\nsiding on disk are converted from ASCII to binary representation, one\nrecord (6-min sequence) for each read instruction.\nA search is then made to locate the Julian date group. The Julian\ndate will be the first three digits with values of 1 59 9 in the record.\nAll other characters (alphabetic) will have a code other than 1 to 9.\nThe time that the sequence was initiated is determined by the identifica-\ntion of a colon (:) between the hour and minute.\nCharacter scanning now identifies station call letters, which are\ncoded as three unique alphabetic characters preceding the data channels.\nThe translated characters are compared with a set of numbers that have\nbeen read in before execution to determine which station call letters have\nbeen identified. The next step after station identification is to check\nby a pre-read code for the station read-in by control cards, what data\nthe station should be transmitting and which channels these data are in.\nIf any characters other than 1 through 9 appear as any digit in the data\nchannel, the channel is considered missing. However, if the character is\nalphabetic, the program checks for another set of call letters. When no\nstation call letters can be identified, the search is continued for the\n32","next data channel of the previously identified station. The three numeric\ndigits from each data channel are packed, checked for values between 1 and\n998, and stored for further use.\nThe u and V wind components, extracted from the data channels, are\nsummed separately until the end of the hour, when the hourly average wind\ndirection and velocity will be computed. Hourly averaged wind direction\nand velocity are computed as follows,\n(1)\n= ((u-500)/N\n(2)\nby use of ATANZ and SORT subroutines from the system library\n(3)\nD = 57.29578*ATANZ(V,U)\nS = SQRT(u**2+T**2)/5.0\n(4)\n.\nThe telemetered values of u and V have been multiplied by 5 for greater\nresolution, and 500 was added to avoid negative numbers. The 6-min inte-\ngrated values of temperature, u, and V are stored in an array during pro-\ncessing, and at the end of each hour the array is written to disk to be\nlater merged with previous data. The hourly averaged wind directions and\nspeeds are printer listed and retained for reference.\nHourly averaged wind directions and speeds are plotted by a Calcomp\nplotter at 3-hr intervals. The rectangular grid is scaled from 220 X\n100 km to 11 X 5 cm. Four maps are plotted for each 22 X 28 cm area of\nplotter paper. The sheets may then be cut to page size. The NRTS site\nboundary is plotted within the grid with date and time labeled above the\ngrid. Wind direction is illustrated by drawing lines from the scaled\nstation location toward the direction from which the wind is blowing.\nBarbs are drawn at 95° perpendicular to the direction shaft for each\n5 m sec-1 of wind speed. Speeds less than 5 are represented by barbs\nof corresponding smaller length.\n4. THE EFFECT OF SAMPLING INTERVAL ON TURBULENT ENERGY SPECTRA\nWhen collecting experimental data, it is advantageous to use as long\na sampling interval as possible in order to reduce the volume of data that\nmust be processed. One must use care in selecting a sampling interval,\nhowever, to insure that the data will be meaningful.\nTo examine the effects of sampling interval on speed and velocity\nspectra calculated from bivane data, we studied data from a bivane located\n4 m above the ground when the wind speed averaged 2.5 m sec-1 . The data\n33","were sampled at an interval of 0.58 sec for a period of 79.2 min; but by\nusing every other data point or every fourth data point, we could compare\ndata with sampling intervals of 1.16 and 2.32 sec, respectively.\nMeans, variances, and spectra were calculated for the wind speed and\nthe three velocity components using all three sampling intervals. For\nthe intervals chosen, no appreciable differences existed in the means or\nthe variances; as can be seen from figures 20 through 25, the speed (S)\nand horizontal velocity components (u and v) are similar except at the high\nfrequency ends where the effects of aliasing can be seen. Our experience\nhas been that good vertical velocity (w) spectra cannot be obtained from\nbivane data, especially when the bivanes are at lower elevations. The W\nspectra calculated in this study are no exception to the rule.\nIf we use the spectra calculated with At = 0.58 sec as a standard,\nwe see that for the U and S spectra aliasing begins to appear at about\n0.3 Hz for At = 1.16 sec, and about 0.07 Hz for At = 2.32 sec. For the V\ncomponent spectra, the effects of aliasing began to appear at about 0.1 Hz\nfor At = 1.16 sec, and about 0.07 Hz for At = 2.32 sec.\nThis study would seem to indicate that unless we were interested in\nthe higher frequencies, the data measured at intervals of 2.32 sec would\nbe almost as good as that measured at 0.58 sec intervals.\n5. SITE EVALUATION PROGRAM\n5.1 Introduction\nA computerized model has been programmed to calculate simulated acci-\ndent and annual average atmospheric dilution factors. This program makes\nit possible to rapidly evaluate potential nuclear power station sites in\na consistent manner.\nFor input to the program, the applicant must provide a joint fre-\nquency distribution of wind speed and stability classes for each direc-\ntional sector from a representative 12-month period. In most cases,\nperiods of calm are included separately as the number of calms occurring\nfor each stability class. The program assigns the calms to the various\nsectors in the same ratio as the lowest wind speed class and adds them to\nthat category. Also included as input to the program are the distances\nto the site boundary for each sector, the cross-sectional area and height\nof the reactor building, and the distance to the nearest population\ncenter. If the reactor has a stack, the applicant also includes the height\nand diameter of the stack and the exit velocity of effluent, the distance\nand bearing of the stack from the reactor, the distances from the stack\nto the site boundaries for each sector, and information on terrain. Ter-\nrain heights and distances are read into the program for each sector;\nthe program performs a linear interpolation to obtain the terrain height\nat any needed location. Plume rise is determined by the formulas devel-\noped by Briggs (1969).\n34","Figure 21. V and W spectra sampling\n4 METER BIVANE SPECTRA. T=79.2 MIN, DT=.58 SEC\n10°\nV\nV\nmust\ninterval = 0.58 sec.\n-1\n10\nCYCLES/SECOND\n-2\n10\n-3\n10\n10\n3\n-2\n2\n-1\n1\n0\n10\n10\n10\n10\n10\n10\nFigure 20. u and S spectra sampling\n4 METER BIVANE SPECTRA, T=79.2 MIN. DT=.58 . SEC\n0\n10\nU\nS\ninterval - 0.58 sec.\n-1\n10\nCYCLES/SECOND\n-2\n10\n-3\n10\n10\n-2\n-3\n0\n3\n2\n1\n10\n10\n10\n10\n10\n10\n10","Figure 23. V and W spectra sampling\nMETER BIVANE SPECTRA, T=79.2 MIN, DT=1.16 SEC\nV\nW\ninterval = 1.16 sec.\n10 -1\nCYCLES/SECOND\n-2\n10\n-3\n10\nif\n-4\n10\n-1\n-2\n3\n2\n1\n0\n10\n10\n10\n10\n10\n10\nFigure 22. u and S spectra sampling\n4 METER BIVANE SPECTRA, T=79.2 MIN, DT=1.16 SEC\nU\nS\ninterval = 1.16 sec.\n-1\n10\nCYCLES/SECOND\n-2\n10\n-3\n10\n-4\n10\n-2\n-1\n0\n2\n3\n1\n10\n10\n10\n10\n10\n10","Figure 25. V and W spectra sampling\n4 METER BIVANE SPECTRA. T=79.2 MIN, DT=2.32 SEC\ninterval = 2.32 sec.\nV\nV\n10 -1\nCYCLES/SECOND\n-2\n10\n-3\n10\n-4\n10\n-1\n10 2\n3\n0\n1\n10\n10\n10\n10\nFigure 24. u and S spectra sampling\n4 METER BIVANE SPECTRA. T=79.2 MIN, DT=2.32 SEC\ninterval = 2.32 sec.\nS\n10 -1\nCYCLES/SECOND\n10 -2\n10 -3\n10\n-2\n10°\n10 3\n2\n1\n10\n10\n10\n10","5.2 Design Basis Accident Model\nAccident meteorology is defined as the value of x/Q that is exceeded\n5 percent of the time during a year of \"typical\" weather conditions. The\nprogram has the option to make calculations for either a ground-level or\nelevated source. The equations used are as follows:\n+ CA) 1/3U (ground-level source)\n(5)\n,\nx/Q = =-2 /U m TTO O Z (elevated source),\n(6)\nUm the upper limit of the wind speed class,\n= the standard deviations of material in the plume in\noyoz\nthe y and Z directions ,\nC = a constant equal to 0.5 ,\nA = the cross-sectional area of the reactor building ,\nh = stack height + plume rise - terrain elevation.\nFor each combination of wind speed and stability class, x/Q is calcu-\nlated and the results are stored along with the associated probability of\noccurrence as determined from the joint frequency distribution. For a\nground-level source, three methods are used to determine the distance at\nwhich to make the calculations:\n1. The distance from the reactor to the nearest site boundary\n(minimum site boundary).\n2. The distance to the nearest population center (low population zone).\n3. The actual distance to the site boundary for each directional\nsector.\nFor an elevated source, the problem becomes somewhat more complicated.\nsame three procedures are used for determining distances, except that\nThe\nthey are measured from the stack rather than the reactor; furthermore, since\nthe maximum value of x/Q may occur beyond the given distance, a searching\nprocedure finds the maximum value of x/Q at or beyond the given distance.\nThe array of x/Q values is ordered from the greatest to the least,\nand the cumulative probabilities associated with them are summed. A\ncurve then is fit to the data points by the least-squares method. A plot-\nting routine has been developed to plot the data points and curve on a\nlog-probability graph (fig. 26) so that both graphical and tabular forms\nof output are possible.\n38","5.3 Average Annual Model\nIn this model, a long-term continuous release is assumed. The pro-\ngram may optionally calculate a ground level, elevated, or mixed release.\nThe basic equation used is\n(7)\nU = the average wind speed in the ith wind speed category ,\nozj = the standard deviation of plume material in the jth\nstability category,\nfijk = the joint probability,\nhijk = the effective stack height,\nC = a constant equal to 0.5,\nD 2\n= the height of the building,\nr = the distance\n= sector width in radians.\n0\nValues of (x/Q) are calculated for the site boundary in each sector\nand for a number of distances out to 80 km. In making use of these re-\nsults, however, keep in mind the uncertainty of values calculated at these\ndistances. Each sector is divided into a number of segments, and an\naverage value of (x/Q) is calculated for each segment.\nThe values of (x/Q) average for each segment may be further smoothed\nby averaging them with neighboring sectors. If desired, we may also ob-\ntain segment averages over all segments at a given distance from the re-\nactor. Output from the annual average portion of the program is both\ntabular and graphical. The graphs include plots of (x/Q) average versus\ndistances for each sector (fig. 27) and contour maps depicting isopleths of\n(x/Q) average surrounding the site area (fig. 28).\n39","10-3\n10\n10-5\n10-6\n10-7\n10-\n9.99\n60\n80\n90\n95\n99\n99.9\n0.01\n0.1\n1\n5\n10\n20\n40\nCUMULATIVE FREQUENCY\nFORKED RIVER DIRECTION DEPENDENT CASE FOR 1968. (DELTA T BETWEEN 200 AND 75 FT.)\nDesign basis accident plot of W/Q versus\nFigure 26.\ncumulative probability for the Forked River\nNuclear Power Station.\n40","FORKED RIVER 1968 (DELTA T BETWEEN 200 AND 75 FT., WIND AT 75 FT.)\n10-5\n10-9\n10\n10\n0.0 DEGREES\n22.5 DEGREES\n10-7\n10\nFROM SITE\nFROM SITE\n10\n10\n9\n10\n10\n0\nI\n102\n10\n0\n102\n10\n10'\n10°\n10'\nDISTANCE IN MILES\nDISTANCE IN MILES\n10-5\n10-5\n10\n10\n45.0 DEGREES\n67.5 DEGREES\n10-7\n10-7\nFROM SITE\nFROM SITE\n10\n10\n10\n10\n10-1\n10°\n0\n1\n10\n10\n0\nI\n102 2\n10\n10°\n10'\nDISTANCE IN MILES\nDISTANCE IN MILES\nFigure 27. Examples of annual average plots of X/Q\nversus distance for each directional sector for\nthe Forked River site.\n41","-9\n5*10\n<<6\n10\n20\n30\n40\n50 MILES\n10\nA\n5x10\n-8\n2X10\n-8\n10-8\n-8\n0\n5x10-9\nCONTOURS OF ANNUAL AVERAGE CHI/Q (SEC/METER CUBED)\nFORKED RIVER 1968 (DELTA T BETWEEN 200 AND 75 FT. , WIND AT 75 FT )\nFigure 28. Contour map of annual average X/Q\nfor the Forked River site.\n42","6. DIRECTIONAL WIND SHEAR UNDER LOW WIND SPEED CONDITIONS\nAn interesting phenomena noted during the initial Building Wake\nStudies was the frequent occurrence of strong directional wind shear in\nthe vertical. During conditions of a stable temperature, lapse rate, and\nlow wind speeds, the winds at the 4-m level of the Grid III tower were\noften flowing in almost the opposite direction from those at the 61-m\nlevel. An example of this phenomena (fig. 29) shows smoke being released\nat three different levels on the Grid III tower. At each level, the wind\nis blowing in a different direction. Figure 30 is an episode that occurred\nthe morning of July 8, 1971, in which the long vectors represent the winds\nat the 61-m level, and the short vectors are the winds at the 4-m level.\nImmediately above each pair of vectors are two numbers. The lower number\nrepresents Mountain Daylight Time (MDT), and the other is the temperature\ngradient (°F/100 ft) as measured between the 61- and 4-m levels. This\nfigure, then, is a series showing wind shear at 10-min intervals with the\ncorresponding temperature lapse rate printed for each plot. The wind\nspeeds at the top of the tower varied from 3 to 6 m sec-1 and at the\nbottom from 0.5 to 3 m sec-1 From 0555 to 0735 MDT, stable condi-\ntions existed. The wind at the top of the tower was fairly steady from\nFigure 29. Aerial photograph showing smoke being\nreleased from the surface, 25 m, and 61 m on the\nGrid III tower.\n43","out of the north or northeast at about 5 to 6 m sec-1, while at the bottom\nthe wind was meandering at about 1 m sec-1. From 0635 to 0655 MDT and\nagain at 0735, the directional shear was about 180°. After 0745 MDT,\nlapse conditions existed and the mixing from above brought the lower level\nwinds more in line with those at the top of the tower. Under these stable\nconditions, the air apparently became stratified and the lower level winds\nbecame uncoupled from those farther up the tower.\nWhen examining the topography of the upper Snake River Valley, we\nfind the plain rises from the SW to the NE. Near the Grid III tower,\nhowever, there is a smaller scale slope in the opposite direction. On\nthis morning, we were probably witnessing a downslope wind at the top of\nthe tower that was governed by the large-scale valley slope, while the winds\nat the bottom of the tower were following the smaller scale slope on a\nlocal level.\nStudies comparing speed spectra with u and V spectra from the upper\nand lower levels on the nearby CFA tower have also demonstrated this un-\ncoupling effect; they showed a diurnal wind reversal at the top of the\ntower that did not exist at the bottom.\nTo determine how frequent and under what conditions this phenomena\noccurs, we instrumented the 61-m tower at Grid III to measure wind velocity\n(10-sec intervals) at the 4, 16, 32, and 61 m levels and temperature\n(1-min intervals) at the 1, 2, 4, 8, 16, 32, and 64 m levels. The data.\ntaken for about 3 weeks, have been stored on magnetic tape, and 10-min\nmeans and standard deviations have been computed. The 10-min average\ntemperature gradients have been plotted for each 10-min period, and the\nwind speeds and directions have been superimposed on the plots to aid in\nthe selection of data cases. Further analysis of the data has not yet\ntaken place, and at this point all we can say is that under stable at-\nmospheric conditions the winds near the surface sometimes become uncoupled\nfrom those above. The character of the diffusion in the different layers\nwould obviously be very different under these conditions. A study of\nthe meandering nature of the surface wind during low wind speed condi-\ntions would probably contribute to the explanation of low wind speed\ndiffusion.\n7. APPLICABILITY OF GEOSTROPHIC WINDS TO MESOSCALE TRANSPORT\nThere are two primary motivations for evaluating the surface geo-\nstrophic winds. First, we are interested in determining transport statis-\ntics by using winds dynamically derived from the historical series of\nmeteorological charts and data. If such an approach should have any merit,\nit would certainly be true at a Plains location such as Oklahoma City.\nSecond, we are interested in determining which elevation above ground\nthe geostrophic winds might best be applied. The presence of the 458 m\nWKY tower in this study serves the purpose.\n44","4.999\n7.925\n7.783\n7.762\n6.523\n615\n625\n635\n555\n605\n3.373\n4.572\n0.081\n4.958\n1.687\n645\n655\n705\n715\n725\nFigure 30. Directional wind\n-1.382\n-1.118\nshear in time.\n-2.215\n-3.333\n-3.658\n735\n745\n755\n805\n815\n-4.328\n-3.231\n-3.962\n-4.288\n-4.511\n855\n905\n825\n835\n845\n-4.531\n-4.816\n915\n925\nBecause of the much denser network of airway observation stations,\nthe geostrophic winds were calculated from observed altimeter settings\nand temperatures rather than computing them from radiosonde observations\nof the geopotential of a near-surface constant pressure level. Using the\naltimeter correction system of Bellamy (1945), we may write the geo-\nstrophic component equations as\n(8)\naz\n(9)\n,\nwhere D is the height difference from sea level of the standard atmospheric\naltitude corresponding to the observed altimeter setting, Z is the terrain\nheight, and\nTVT\n(10)\nS*\nT\np\n45","where T is the observed virtual temperature, and T is the temperature\ncorresponding to the station pressure altitude in p the standard atmos-\nphere.\nThe 1° latitude by 1-1/4° longitude computational grid, and locations\nand altitudes of the 10 observing stations are shown in figure 31 and\ntable 5. D, Z, and S* values were interpolated to grid points according\nto the inverse-square weighting of the observations of the corresponding\nvariables evaluated at the nearest four stations.\nBecause of the critical dependence of the derived winds upon the com-\nputed gradients, the SJ scores (Tweles and Wobus, 1954) were computed for\nthe inverse square and inverse linear derived gradients of D against a\ncareful hand analysis, where\n(F-)\nS1 =\n/n ,\n(11)\nMAX(F,O)\nn is the number of gradient measurements on a map, F is the gradient from\nthe objective analysis, 0 is the gradient from the hand analysis, and\nMAX(F,) is the maximum of the values F and O. Scores were computed for\nfour map times, 0200 and 1300 CDT of both October 1 and October 9. There\nwas no statistically significant difference between the inverse square and\ninverse linear methods of objective analysis. The S scores were all with-\nin 0.305 to 0.435, and the maximum difference between inverse square and\ninverse linear for any one map time was 0.018. At the interior of the\ngrid, and particularly the grid points surrounding Oklahoma City, the\ndifferences between objectively and subjectively derived gradients,\nFigure 31. Computational grid for\nPNC\ngeostrophic winds in Oklahoma.\nGAG\nTUL\nEND\nThe grid squares are 1° latitude\nby 1.25° longitude. The numbers\n5\nare identifiers of gradients.\n19\n18\nOKC\n9\nHBR\nFSI\nMLC\nADM\nSPS\n46","Table 5. The Stations , Observations the Grid Point Values\nof D,2, and S* were Interpolated From.\nStation Name\nLatitude\nLongitude\nAltitude\nCoordinate\n(°N)\n(°N)\n(CM)\nX\ny\nGage, Okla. (GAG)\n36.30\n99.77\n680\n3.42\n2.30\nEnid, Okla. (END)\n36.38\n97.80\n360\n1.84\n2.38\nPonca City, Okla. (PNC)\n36.73\n97.10\n308\n1.28\n2.73\nTulsa, Okla. (TUL)\n36.20\n95.90\n206\n0.32\n2.20\nOklahoma City,\nOkla. (OKC)\n35.40\n97.60\n392\n1.67\n1.40\nHobart, Okla. (HBR)\n35.00\n99.05\n477\n3.84\n1.00\nFort Sill, Okla.\n(FSI)\n34.60\n98.40\n362\n2.32\n0.60\nArdmore, Okla. (ADM)\n34.30\n97.02\n238\n1.22\n0.30\nMcAllestor, Tex.\n(MLC)\n34.88\n95.78\n229\n0.22\n0.80\nWichita Falls, Tex.\n(SPS)\n33.97\n98.48\n310\n2.\n.38\n-0.03\naz\naz\nand\n, was generally less than 3 m per grid interval (approximately\nax\nay\n108 km).\nThe initial test was conducted by evaluating the geostrophic wind at\nOklahoma City by interpolation of U determined from gradients 18 and\n19\nand V determined from gradients 5 g and 9; figure 31 shows the referenced\ng\ngradients. The winds observed on the WKY tower and the geostrophic winds\nderived by the inverse square interpolation are shown in table 6. Differ-\nence between the geostrophic winds and the OKC surface winds is depicted\nin figure 32.\nIn the south wind situation of October 1, the nocturnal low-level\njet was producing supergeostrophic winds after 0600 CDT at the middle level,\nand after 1000 CDT at the top of the tower. On October 9, a cold front\npassed the WKY tower shortly after 0400 CDT. This front was responsible\nfor considerable mesoscale departure of the local pressure gradients at\nthe WKY tower as opposed to those derived using Oklahoma City and Ponca\nCity to the north. (Enid was missing until 0900 CDT.) The directional\ndifference between the geostrophic wind is generally 30 to 40° in the south\n47","Table 6. Observed Winds on the Hour at Three Levels of the\nWKY Tower and the Computed Geostrophic Wind.\nTime\nWKY\nTower\nGeostrophic\n(CDT)\n90 m\n266 m\n450 m\nWind\nOctober 1, 1971\n0200\n1715*\n1835\n1742\n2020\n0300\n1817\n1835\n1843\n2121\n0400\n1819\n1831\n1842\n2123\n0500\n1815\n1836\n1842\n2025\n0600\n1816\n1934\n1942\n2232\n0700\n1716\n1928\n1934\n2233\n0800\n1815\n1828\n1944\n2235\n0900\n1817\n1830\n1935\n2234\n1000\n1918\n1827\n1936\n2235\n1100\n1920\n1825\n1929\n2139\n1200\n1819\n1923\n1925\n2239\n1300\n1720\n1924\n1822\n2238\nOctober 9, 1971\n0200\n0000\n0102\n3603\n0113\n0300\n0000\n3402\n3307\n0115\n0400\n2603\n3006\n3211\n0318\n0500\n3112\n3216\n3218\n0422\n0600\n3413\n0323\n0118\n0322\n0700\n3512\n0223\n0122\n0425\n0800\n3412\n0221\n0119\n0424\n0900\n3414\n0220\n3621\n0328\n1000\n3411\n0124\n3619\n0221\n1100\n3510\n0123\n0119\n0221\n1200\n3609\n0111\n3612\n0119\n1300\n3514\n3515\n3514\n0117\n* Direction, 170° and speed, 15 kt.\nwind case, and generally 00 to 20° in the north wind case when comparing\nwith the observed winds at the 266-m level of the tower.\nDuring the period of the low-level nocturnal jet, the lowest level of\nthe tower was in best agreement with the geostrophic winds, otherwise the\ntop level was in best agreement during the south wind period. There was\nno significant difference between the agreements of the middle and top\nlevels with computed surface geostrophic winds for the north wind case.\n48","DIRECTION ERROR\n+100\n187\n+0\n-100\nVector error IVI\nFigure 32. Difference of com-\n+40\nputed geostrophic wind from\nobserved wind at Oklahoma\nCity (V - V ). Solid line\nObs\ng\ndashed line -\nOct.\n1;\n20\nOct. 9.\n0\n4\n8\n12\nHour (CDT)\nCertainly the data sample is very small, consisting of only two ran-\ndom cases, therefore, conclusions must be drawn with reservation. These\nresults show large differences between geostrophic and observed winds and\ntend to confirm the intuitive suspicion that derived geostrophic winds are\ngreatly different from derived mesoscale windfields for determining trans-\nport of material in the atmosphere. An important consideration in deriving\ntransport is the grid interval of computation. In the mesoscale windfield\nstudies, the interval was 8 km whereas, the interval for geostrophic winds was\napproximately 110 km. The large computational interval for the geostrophic\nwinds precludes determining fine detail of the characteristic transport in the\nmesoscale. The differences in direction and speed of geostrophic winds when\ncompared with the observed winds are almost certainly greater than the same\ndifferences with the oscale-derived - windfield, even for grid points that\nare in juxtaposition. For the effort involved, and the degree of accuracy\nattainable, computed geostrophic winds appear to be an unadvisable approach\nto determining transport of material over a region only a few tens of kilom-\neters square. Actual wind records from a single station, though inferior\nto a windfield derived from many wind stations, are of at least equal value\nto computed geostrophic winds in determining transport in the mesoscale.\n49","8. AN APPLICATION OF THE MESOSCALE WINDFIELD METHOD\nAn historical data set of hourly wind records from several stations in\na mesoscale area may be used to answer a number of questions. As an ex-\nample, it may be desirable to determine how often air may move from a\nsource 'S' to a small receptor area 'B' in 6 hr. At the NRTS, it was neces-\nsary to know how often mid-September daytime winds would move air northeast-\nward from the CPP facility so that it would cross each of four 30° tracer-\nsampling arcs (the outermost arc being 100 km from the source). This must\ncontinue to do so for 6 hr and end by sunset. It was also necessary to\nknow whether such successful cases are predictable 12 to 36 hr in advance\nfor tracer-test planning, and whether prediction with almost certainty\n1 hr in advance was possible for the initiation of the tracer release.\nFor the period of concern (September 11 to 22), frequencies of success-\nful cases were determined by plotting daytime hourly trajectories for a\n3-year sample for September 10 to 25. The first part of the 1969 sample\nis shown in figure 33. There were seven successful cases of a possible 48\nin the 3-year sample.\nWhether or not these successful cases could be predicted on the plan-\nning scale, that is, 12 to 36 hr in advance, required an examination of the\nsynoptic weather charts associated with each case. Figure 34 is a compo-\nsite of the 1200 GMT 500-mb charts on the days of success. The 41 remain-\ning 500-mb charts were then scanned to see if any were highly correlated\nto the composite. This would not be a reasonable thing to do in every such\nexperiment; however, the seven charts that were composited showed amazingly\nlittle diversity between them, and all appeared to belong to a definite\npattern typified by the composite. The pattern was basically a zonal flow\nwith the trough line somewhere between 120°W and 130°W. Five cases were\nfound where the 500-mb chart was well correlated (by visual inspection) to\nthe composite chart. In fact, the map of September 19, 1970, looks almost\nidentical to the composite chart but was an unsuccessful case. The problem\nis to see whether the surface chart is able to be used as a discriminator\nbetween successful and unsuccessful cases where the 500-mb chart is well\ncorrelated to the composite chart of successful cases. It would consume\nmuch space for discussion and figures showing the 12 surface charts asso-\nciated with the 500-mb charts of the composite 500-mb pattern and those\nthat are well-correlated but belonging to unsuccessful cases. Suffice\nit\nto say, therefore, that pre-frontal; post-frontal; surface low pressure\ncenters to the NE, E, SE, and S; and high pressure moving in from the NW,\nW, and SW were all connected with both the successful and unsuccessful\n12 cases of well-correlated 500-mb charts. It appears that with the use\nof prognostic 500-mb charts, it is possible to give alerts 12 to 36 hr in\nadvance for very nearly all successful cases; however, nearly 50 percent\nof the alerts will not materialize as successful cases.\nThe final step, namely the ability to accurately predict in the last\n1 or 2 hr preceding its onset, whether or not a case will be successful\n50","CPP 1100 09/10/69\nCPP 1100 09/11/69\nCPP 1100 09/12/69\nCPP 1100 09/13/69\nCPP 1100 09/14/69\nCPP 1100 09/15/69\nCPP 1100 09/16/69\nCPP 1100 09/17/69\nCPP 1100 09/18/69\nCPP 1100 09/19/69\nCPP 1100 09/20/69\nCPP 1100 09/21/69\nCPP 1100 09/22/69\nCPP 1100 09/23/69\nCPP 1100 09/24/69\nCPP 1100 09/25/69\n12\nFigure 33. Series of four trajectories beginning from\nCPP at dates and times shown.\n51","becomes essentially an ability to assess the persistence of the existing\nwind condition as shown by the mesoscale windfield.\nOnce the basic SW flow begins, the subsynoptic scale pressure gra-\ndients appear to be quite well correlated to the variations between S and\nW components. If pressures are quite low in SW Idaho relative to the\nupper Snake River Plains, the wind will invariably have too much southerly\ncomponent to cross the arcs, as on September 18, 1969 (fig. 35).\nSeptember 19, 1968, shown in figure 36, is a typical case of too much west-\nerly component because of high pressure to the north. The final deter-\nmination for the 1-hr prediction of success still depends upon having both\na favorable flow pattern and an ability to monitor the mesoscale windfield.\nIn this study, we see the mesoscale windfield being used to determine\nsuccessful and unsuccessful cases in a data sample, and finally being used\nin real time for subjective decisions.\n9. LONG-TERM TRAJECTORIES DETERMINED FROM SINGLE- AND MULTI-STATION MODELS\nConsiderable effort has been expended to standardize the required cal-\nculations of environmental hazards for the safety analysis reports of pro-\nposed nuclear reactors, and to prepare environmental impact statements\nabout industrial sources of a variety of atmospheric pollutants. Though\nmodels using all available wind records do exist, the standard calculations\nuse a wind rose from a single station to determine frequency of transport\nin the various directions from a source. Because of characteristic turn-\nings in the windfield and diurnal changes in the wind's character, the\nadequacy of a single wind record is rather limited. This study compares\nthe long-term, two-dimensional distribution patterns of trajectories de-\nrived by the wind rose with the windfield determined from many stations.\nAn earlier study (NOAA Tech. Memo. ERL ARL-32, 1971) lacked the needed\ndata from a plains station and did not allow recirculation or extended\ndwell of air parcels being moved in the windfield. However, an idea of\nthe divergence between the distribution patterns obtained from single sta-\ntion versus those of multi-station was established.\nSimulated trajectories are derived on a computer from one release\npoint in each of three geographically diverse regions: the NRTS, Los Angeles,\nand Oklahoma City. The method of determining the trajectory is that devel-\noped by Wendell (1972). In each of the three studies, one particle is re-\nleased at the beginning of each hour, and advected for 12 hr, or until the\ntrajectory leaves the grid, whichever comes first. The data samples were\nfor September 1 to October 1, 1969, at Los Angeles; September 21 to\nOctober 12, 1971, at Oklahoma City; and 1969 at the NRTS.\nThe y-x computational grid units were 30 X 20 at the NRTS, 13 X 13 at\nLos Angeles, and 16 X 11 at Oklahoma City. The grid interval was 4.3 km\nat the NRTS, 5 km at Los Angeles, and 8 km at Oklahoma City. Smaller\ndetails may be resolved on smaller grids; however, the disparity in the\n52","WA\n940\no\n069\nIF\nLB\no\n931\n920\n50\n120\n10\nXD\n888\n879\nVG\n552\nRM\nQW\n28\n869\nAE\nHB\nRV\n878\n876\no\na\nBA\n887\n873\n88.\n868\no\nYO\n891\n1\nKY\ny\n22\n55\n30°\nO\nDC\n877\nXX\n5\n885\n886\nQV\nOBLI\n00\nSD\nXC\nO\n200\n889\no\nXH\n138\n884\n89\nYN\n0124\nMJOR\nOL\n45°\n88\n875\no\nXB\no\n82\n89\nO\n874\n53\nSMP\n861\nE\nFCA\nCOF\nGEO\nEN\nHVR\nMLP\nO\no\n785\n862\nCTR\nGFA\n777\nGGW\nEO\nGTF\nWSN\nMO\nALW\nDLS\nLWO\nMSO\no\n76\n570\nO\nO\nONP\nDRU\nPDT.\n775\n767\n773\n83\nHLN\no\n772\nOTH 693\nRDM\n1687\nDIK\nOHIA BZN\nMLS\nBKEO\nO\no\n6790\nRBG\nSMN\nB11\nBO\n691\nOLVM\n0690\nO\nO\nLEM\nSXT\n686\n677\nNO\nOMER\nWEY-\n669 ME\n683\nno\nBO\nCOD\nSHR\n667\nDBSO\n676\nSIY.\n5\n68\nACU\nAT\nIDA\n674\n666\nO/\nRCA\nMHS\nGNG\n-589\n663\nRAP\nPHP\nQ\nBY\nO\n(594\n662\n578\nLND\nBPI\nLD\nCPR\nSVE\n58\nO\nDGW\nFOB\nWMC\nO\nCDR\n576\n591\n577\n569\nP\n1584\nEKO\nOGD\n5900\n583\nLOL\nHIF\nRKS\nRWI\no\nFBR\nBLU\nENV\n582\n0575\nSCT\n80\n574\nO\n581\nLAR\n566\n516\nSAC\n488\nFFN\nTO\n1250\nCYS\nSNY\n572\n0483\nO\nO\n487\n563\nELY\n564\no\nO\nDTA\n570\nIML\n571\n486\nFSR DEN\nAKO\n509\nTPH\nRIL\n479\nEGE\nLRY\no\nMLF\nGJT\n469\n0475\no\nHVE477\no\nGLD\n476\nCDC\n467\ncos\n0\nCLG\no\n389\n465\no\nSGU\nPUB\n387\nBDG\n35\n397\nLHX\nLAA!\n396\no\nLSV\nCEZ\nBF\nNID\n464\nO\nGCI\n472\nLAS\nDRO\nALS\nSMX\no\n463\nTAD\nPGU\n0394\n384\no\nO\nPMD\n386\nGCN\n462\nO\nSDB\n3830\n39\nDAG\nSBABURO382\nGUY\nFLG\n368\nO\nDHT\nOEED\nINW\nZUN\nSAP LVS\nNSI\n295\nPRC 375\n380\n297\nTRM\nO\n374\nABO\nBLH\nTCC\n372\nQOTO\n292\nAMA\n365\nO\nSAN\nELC\n363\nPHX\nCVS\nCHD\nYUM\nGBN\no\n278\n362\nO\nROW\nREE\n005\n280\nTCS\no\nRSW\nLUB\n30\n050\nTUS\n276\nO\nTUO\n268\n267\nPPF\nWSD\nHOB\n274\no\nCUS\n06\nDUG\nBGS\no\nSFI\nO\nMAF\nO\nALO\no\nINK o\nComposite 500-mb chart of 7 days resulting in\nFigure 34.\nfavorable SW winds for testing on the long range\ndiffusion test grids.\n53","0001 966 966\n8001 2001\n2101\n9101\n0201\n68\n99\nfor\nHOIH\n6E\nHSD\n82\n1940\nLE\nKz\nas\n6EI\nBC\nthe\n2LE sigh\nas\nSo\n158\nss\nat\nis\n09/81/60 0011 ddg\nLS\n802\nGoe\nIt\nSD\n9101\nLS\nas\n82\nEgi\nE81\nStoLE\n181\n070\n8D\n9L1\n01\nED\n102\nSS\n8D\n591\n6S\n/E+O\n391\nEb\n19\n10140\n19\nSS\n89\nato is\n19\n88\n000\n696L 77as 'IND 003I of sishquo thof FIDYO fo 70s annfans pub\ndd '8' worf fixp awas 247 COET 70\nSE\nto","1016\n1016\n1020\n102\n1020\n1020\n634174\n2=101\n59.102\n880\n12/03/19\n+104\n40\n101\n39\n085\n+2\n36\nO\nLOW\n39\n1089\n070\n39\nCPP 1300 09/19/68\n-\n48\n48\n28\n36\nso\n065\n46\nLOW\nHIGH\n1008\n63\n50/102\n48\n59\nSS\n082\n48\n48\n122\n53\n52\n075\nFigure 36. Same as fig. 35, except date is Sept. 19, 1968.\n55","computational grids is insufficient to make any significant difference in\nthe basic patterns of the transport statistics.\nTwo basic transport statistics are computed to compare the results of\nthe contemporary \"single-station\" long-term dosage calculations with the\nsame calculations using a windfield for transport. The single-station cal-\nculations of transport statistics are drawn from a wind rose that is com-\nposed of winds averaged over an hour for the period of record, while the\nwindfield transport statistics are drawn from the derived trajectories.\nNote\nthat what is being computed are \"transport statistics\" and not \"dos-\nage statistics, since no time-dependent function of dosage rate was used\nonce the hypothetical particles were released.\nTransport statistics using a wind rose are computed two ways. The\nfirst, called \"wind rose transport statistics,\" is to simply determine\nthe proportions of each direction ray that are intercepted by each grid\nsquare, and multiply the nonzero proportions by the frequencies for those\nrays. Figure 37 shows a typical grid square intercepting parts of two\ndirection rays, about 50 percent of the 240° ray, which includes angles\nfrom 235° to 245° and about 40 percent of the 250° ray, which includes\nangles from 245° to 255°.\nThe frequencies are adjusted to account for winds too weak to advect\nparticles as far out as a given grid square within the given time limit.\nFor the advection time limit of 12 hr, a grid square whose center is more\nthan 22 km from the source will not include 0.5 m sec-1 winds in the wind\nrose frequencies.\nBefore discussing the second approach to using the wind rose, the\n\"windfield transport statistics\" method should be explained. The wind-\nfield derived trajectories were interpolated for each 10-min time step\nexactly as in the studies already cited. For each 10-min segment of\neach trajectory, each grid square crossed or entered was given an incre-\nmental count. This method is only a step from computing dosage, because\nthe final statistics reflect the total dwell time of trajectories over\ngrid squares.\n240\nFigure 37. Grid square inter-\n250°\ncepting parts of two direction\nrays of a wind rose.\nS\n56","The second wind rose method is analogous to the windfield method in\nthat it considers the total dwell time of a trajectory over a grid square.\nThe wind rose was an s x d matrix, WR where 's' is the number of speed\nclasses and 'd' is the number of direction classes. The latter was 36 at\nthe NRTS and Oklahoma City, but only 16 at Los Angeles. The number of\nspeed classes was taken as 40, a class for each unit increment of speed in\nmiles per hour up to 40 mph. All the cases belonging to a speed and\ndirection class were advected simultaneously in 10-min steps, as with the\nwindfield trajectories. To get a smoother contour pattern, the trajector-\nies of each direction class were divided evenly over 1° sub-intervals and\nadvections made for each of the 360 sub-intervals. Trajectory frequency\nwas counted the same way as with the windfield. The method shall be\ncalled \"wind rose - total hits analogy.\"\nThere is a boundary problem with the windfield computed statistics\nthat results in an unfavorable comparison to wind rose computed statis-\ntics. The nearer the boundary, the larger the number of trajectories that\nin reality should have been recirculated over the grid, but could not be\nbecause they had been advected off the grid before the recirculation began.\nThis results in a negative bias of counts in the windfield computed statis-\ntics; this is least near the center of the grid but may be very sizable\nnear the boundaries, particularly if diurnal wind shifts are characteristic.\nFor each of the three geographical areas, five matrices of transport\nstatistics were derived; windfield (WF), wind rose (WR), wind rose - total\nhits analogy (WRTHA), the ratio WR/WF, and the ratio WRTHA/WF. The two\nratio matrices were normalized and smoothed before entering the computer\ncontour plotting routine. Normalizing was accomplished by multiplying\nall matrix elements by the inverse of the value of the matrix element\nrepresenting the pollutant source. The smoothing was accomplished with the\noperator of = + av20 (Dingle and Young, 1965). By repeated applications\nof different values of a, a very selective damping of short wave \"noise\"\nand nonamplification of longer waves may be accomplished. For this test,\na was set equal to 1/8 and given only one pass. The resulting damping\ncurve is shown in figure 38.\n1\nFigure 38. The damping curve for\n= 1/8 in the operator\n0\n1\n2\n3\n4\n5\n6\n7\n8\n9\n10\nWave length in grid units\n-1\n57","The transport statistics matrices were then computer contoured with\nan overlying geography to give orientation and geographical references.\nThe dimensions of the borders are 96 X 132 km at Oklahoma City, 68 X 68\nkm at Los Angeles, and 45 X 135 km at the NRTS.\nThe contoured patterns are shown in figures 39 to 41. Oklahoma City,\ncentered on the WKY tower, is shown first because of the relatively good\nB - WRTHA (1213)\nA WR (521)\nC WF (1203)\nSTW\nSTW\nSTW\nHEN\nHEN\nHEN\n2\n2\nGTH\n1\n2\nGTH\nGTH\n3\n3\n5/\nHERY\nWKA\nMKR\nMKR\nMKR\nELR\nELR\nELR\nMOR\nMOR\nMOR\nCHK\nCHK\nCHK\nWAN\nWAN\nWAN\nD WR/WF\nE - WR THA/WF\nSTW\nSTW\nHEN\nHEN\n3\nGTH\nGTH\n2\n3\nWKY\n.\nWKY\n1\nMKR\nMKR\nELR\nELR\n2\nMOR\nMOR\nCHK\nWAN\nWAN\nFigure 39. Total hits analysis for Oklahoma City. Advection is by\nWR - wind rose; WRTHA - wind rose total hits analogy; and WF -\nwindfield. Panels D and E are the indicated ratios. The numbers\nin parentheses in panels A, B, and C are the values for the release\nsquare. The contours represent 5 to 25 percent in 5 percent incre-\nments of the release square value.\n58","A - WR (8122)\nB - WRTHA (14964)\nC - WF (22641)\n1\n2\nD WR/WF\nE - WR THA/WF\n2\n.\n34\n4\nFigure 40. Same as fig. 30, except for release from LOFT at the NRTS.\ncorrelation between patterns derived by single-station (wind rose) records\nand patterns derived by multiple station (windfield) records. The NRTS\n(fig. 40) centered on LOFT in the north, shows quite good correlation of\nsingle-station and multi-station results except for certain areas of charac-\nteristic events in the mesoscale windfield. The final set (fig. 41) shows\nthe Los Angeles results centered on \"F\" near Long Beach where correlations\nof single-station and multi-station transport statistics are extremely poor.\nIf one inspects the trajectory patterns for an entire period of record,\nas done at the NRTS (Wendell, 1970), it is surprising that with the large\nnumber of individual cases that show characteristic turnings and recircula-\ntion, the wind rose transport patterns compare with windfield transport\npatterns as well as they do. Inspection of the daily sets of trajectories\nat Oklahoma City revealed a large percentage of days when trajectories had\n59","A WR (765)\nB - WRTHA (4939)\nC - WF (3356)\nBUR\nBUR\nBUR\n12\n1\nWHIT\nWHIT\nWHI\n2\n3\nANA\nANA\nANA\n3\nF\nD WR/WF\nE - WR THA/WF\nBUR\nBUR\n5\nWHIT\n4\nANA\nANA\nFigure 41. Same as fig. 30, except for release from\n'point F' at Los Angeles.\ncurved patterns rather than straight. Apparently a strong tendency exists\nfor veering, backing, or even direction-reversing - trajectories to be re-\nplaced along their initial straight-1 ine path by trajectories that started\nin other directions. Consequently, though large numbers of individual case\nstudies show complex patterns of trajectory meanderings, there is a long-\nterm statistical averaging of trajectories that reflects only character-\nistic perturbations in the mesoscale windfield.\nAt Oklahoma City, the model wind directions are SSW and NNW with nearly\nnormal frequency distributions in the SSE through SW, and NNE through NNW.\nThere are minor maximums for ESE and W winds. The southerly winds have the\ngreatest average velocity, as noted by the shrunken contours to the north\nof the WKY tower on the WRTHA plot as compared with the WR plot. These\nshrunken contours would indicate either less dwell time, or higher velo-\ncities. The windfield plot shows the maximums at the north and south bor-\nders to be farther east than with the wind rose plots. A significant\nmaximum also exists at the border directly east from WKY. This maximum is\nonly vaguely suggested on the wind rose plots. The frequency minimum to\n60","the NE of WKY on the windfield plot would indicate a tendency for SW winds\nto lack persistency, and shift more to southerly at times, or to shift from\nS through W to N with front passages.\nThe ratio patterns reveal how well the single-station patterns depict\nthe windfield patterns. The WRTHA/WF ratio is the best of the two ratio\npatterns by reason of the use of \"dwell time\" in the computations of WRTHA.\nRatios ranged from 80 to 120 percent within about 11 km of the WKY tower\nrelease point. Along the directions of primary flow, the ratios remained\nwithin 20 percent of unity all the way to the border. In the easterly\ndirections from WKY, the ratios fell to as low as 40 percent at 16 km. This\narea was affected by several wind reversals. It means that at one point\nabout 16 km east of WKY, the actual probability of hit or actual long-term\ndosage is at least twice as often as would be determined from a wind rose.\nThe NRTS transport frequency patterns show two particularly interest-\ning features. The range of directions for N to E winds at LOFT is fairly\nuniform and results in a well-spread WRTHA pattern in the SW part of the\ngrid. In reality, a strong channeling tendency exists that yields the\nnarrower pattern seen on the windfield plot. The WRTHA/WF ratios are only\n45 to 75 percent along the axis of the actual flow and become 150 to 200\npercent - 8.24 km either side of the axis from the south part of the NRTS\nand continuing south. The second area of special significance is east of\nthe NRTS boundary, at the latitude of LOFT, which is a known stagnation\narea in the middle of a converging counterclockwise circulation, which is\ntypified by the windfield series shown in figure 42. The area affected\nis in the direction of a minimum in wind rose frequencies, while the stag-\nnation effect is clearly shown in the windfield pattern as a pronounced\nmaximum. The area about 24 km directly east of LOFT is characterized by\nWRTHA/WF ratios on the order of 33 to 45 percent. The area of ratios less\nthan two-thirds extends nearly 65 km east of LOFT.\nThe diurnal wind reversals at the NRTS occur with such frequency that\na major proportion of trajectories crossing any part of the upper Snake\nRiver Plain are what we call recirculated trajectories; that is, they\nhave experienced a wind reversal. Hence, the boundary effect, which was\ndiscussed earlier, is especially critical at the NRTS and explains the\nrapid change in the ratios near the boundaries, especially the eastern\nand western boundaries which are not along the primary axis of flow.\nThe Los Angeles study is unique in that the chosen release point 'F'\nis 11 km WSW from the nearest wind station. In a plains location like\nOklahoma City, this remoteness of the wind station from the release point\nwould not be so serious so that wind rose frequency statistics could be\ncalculated as though wind station and release point were in juxtaposition.\nAt Los Angeles, it depends upon whether or not the remoteness means the\ntwo positions are in differently oriented parts of a characteristically\ncurving flow. The release point 'F' is in the eddy behind the Palos\nVerdes hills that lie to the west. The characteristic on-shore flow at\npoint 'F' is from the WNW-WSW, while the wind station 11 km to the ENE is\n61","2200 01/08/69\n2300 01/08/69\n2000 ou 08/69\n2100 01 /08/69\n1200 03/06/69\n1300 03/06/69\n1400 03/06/69\n1500 03/06/69\nFigure 42. Series of windfield plots showing formation 1969.\nof large eddy, 2000 through 2300 MST, Feb. 9,\nSSW sea breeze. The resulting wind rose is not representative the ratios of\nin the prevailing winds at the release point, and consequently large or of\nthe and windfield statistics are almost exclusively misleading very even\nwind rose and the wind rose transport statistics are is an excellent\nvery small, the first kilometer from the release point. This\nwithin of using an available wind record for a transport frequency one.\ndemonstration study that is not only a poor guide, but also a completely misleading\nsome important conclusions that may be drawn from the the NRTS, inves-\ntigations There of are wind rose versus windfield transport statistics at\nOklahoma City, and Los Angeles.\nIn real accident or emergency situation, the use of a single-\n1.\nstation a wind record for distances greater than those involving\nevacuation should be in the hands of personnel experienced the\nsite in boundary layer and small-scale flow. This is because\npollutant is being transported by a turbulent fluid with vertical\nas well as horizontal inhomogeneities and perturbations.\n2. For estimates of maximum probability of hit in event of an\naccident, and for estimates of maximum long-term dosage, the\nwind rose may be a reliable indicator for distances to something\nlike 80 km, providing no geographical features are encountered\n62","that would induce characteristic changes in wind direction. There\nis likely to be a slight difference in direction between reality\nand the wind rose, as was the case at Oklahoma City, so that at\n80 km the uncertainty in the position of the maximum is 32. km\nin either direction along the arc; 32 km represents about 20°\nin direction from the source at the 84-km arc.\n3. For general estimates of probabilities of hit during an accident\nand for long-term dosages, there are so many uncertainties in geo-\ngraphical effects and characteristic small-scale meteorological\nfeatures that reliable distributions may be estimated to only\nsomething like 24 km at a plains station, and only a couple of\nkilometers for a station near a barrier or seashore, which\ncould cause great horizontal discontinuities in wind direction.\nThese studies do establish a need for caution in taking single-\nstation wind records to determine long-term patterns of probabilities of\nhit and dosages.\n10. LONG DISTANCE TRANSPORT AND DIFFUSION TO 100 KM\nA need exists to understand transport, diffusion, and plume depletion\neffects upon airborne material carried out to about 100 km and lasting for\nmany hours. Extensive diffusion measurements have been painstakingly made\nin the past couple of decades and have been summarized and reported by\nmany authors (e.g., Heffter, 1965; Slade, 1968). In general, these mea-\nsurements have described diffusion at the shorter distances or have dealt\nwith large instantaneous sources at long distances, high altitudes, or\nboth. The intent of our study was to improve understanding of the trans-\nport and diffusion of quasi-continuous effluent releases dispersed within\nthe planetary boundary layer.\nAt the initial stages of such a measurement program, the most reliable\ncourse of action seemed to be one that would include several simultaneous\ntypes of measurements (a) to determine the consistency of the different\nmeasurement techniques, (b) to combine the complementary data types to gain\na more complete overview of the effluent behavior, and (c) to identify and\nemphasize data measurements that appeared to be the most vital ones. At\nthe beginning of such a study, it is crucial to directly measure effluent\nconcentration in order to evaluate the effective plume trajectory and dilu-\ntion. The key information available from this testing was the measurement\nof tracer ground-level concentrations of successive distances downwind.\nBased on these measurements, the downwind change of relative axial concen-\ntration and lateral tracer dispersion could be examined. Along with these\ndirect tracer measurements, this testing included unique, simultaneous\nmeasurements of low-level winds from an encompassing network of tower-\nmounted wind sensors and trajectories from serially released tetroons.\nThe information presented illustrates the successful field implementation\nof a unique combination of complementary and valuable measurements that\n63","could be a useful technique for achieving a better understanding of\nregional dispersion of airborne material.\nTwo field tests have been conducted to date. The first test (NOAA\nTech. Memo. ERL ARL-32, 1971 was of limited use because of the presence\nof a wind shear or discontinuity zone across the sampling grid some 40 km\ndownwind. The second test, conducted on August 31, 1971, was highly suc-\ncessful and illustrates comparisons of the various transport and diffusion\nestimation techniques. A third test was aborted due to the failure to\nachieve suitable testing conditions.\nThe measurements described in this study were conducted over the\nUpper Snake River Plain in southeastern Idaho at the NRTS and adjacent\nareas. The Upper Snake River Plain is approximately 1500 m above sea level\nand is enclosed by mountains that rise 1200 to 1800 m above the valley floor\nalong all but the southwestern edge of the rectangular study area. The\nirregular boundary in figure 43 outlines the NRTS.\nThe tracer used in this investigation was methyl iodide gas labeled\nwith the radioisotope iodine-131. This technique was developed and used\nduring a limited number of diffusion comparisons with uranine dye and\nmolecular iodine gas for downwind travel distances less than 3.2 km\n(Start, 1970). This gas is chemically inert, for practical purposes, and\nallows a control of plume depletion due to dry deposition. Sampling checks\nhave not revealed a tendency for ultraviolet decomposition of the methyl\niodide into reactive (depositing) iodine species. Time integrated air\nN\nARC D\nFigure 43. Field measurement area\nARC C\nwith wind tower locations (dots)\nand tracer sampling arcs (A,B,\nC, and D). The tracer release\nlocation is designated ICPP.\nARC B\nARC A\nICPP\n16 MILES\n64","samples were collected by a network of high volume air samplers, each\nloaded with an iodine impregnated charcoal cartridge and a glass fiber\ntype prefilter. The sampler collection efficiency for methyl iodide was\n0.75. Four sampling arcs, with from five to eight high volume air samp-\nlers per arc, were located at approximately 6, 19, 48, and 80 km down-\nwind. Iodine-131 in the samples was qualitatively and quantitatively\nidentified using gamma spectroscopy and conventional well scintillation\ncounting techniques. Methyl iodide gas was released for 65 min from the\n76-m stack at the Idaho Chemical Processing Plant.\nTwenty-four tower-mounted wind sensors were operated within the rec-\ntangular area of figure 43. Wind directions and speeds measured at some-\nwhat randomly located tower sites within the grid were interpolated to a\nrectangular mesh of points. Calculated trajectories were derived from\nthese \"windfields\" for comparison with observed plume centerline positions\nand tetroon trajectories.\nA transponder-equipped tetroon was released about 1 12 hr before the\nmethyl iodide release. Other such tetroons were launched at the beginning\nof the release and at approximately 20-min intervals during the ensuing\nhour. In all, five tetroons were launched. The tetroon-transponder\nflights were tracked by a M-33 radar. Since a single radar was following\nseveral tetroons at the same time, positioning of individual tetroons was\naccomplished about once every 5 min. Detailed tracking information was\nnot determined from any one tetroon; rather; the general position of each\none was determined so that the lateral and vertical envelopes of their\npositions for the series of launches could be estimated. The tetroons re-\nmained within 600 m of the surface much of the time. All tetroon posi-\ntions during vertical height oscillations dipped to within 1600 m of ground\nlevel. The Boise radiosonde data and the 700 and 500 mb analyses for 1200\nGMT on August 31 and 0000 GMT on September 1, 1971, were examined. From\nthis information and the surface temperatures in the test area, a maximum\nadiabatic mixing depth of 1400 m was estimated. This value was similar\nto the 1600-m depth that contained the tetroon oscillations.\nNormalized axial concentrations versus downwind distance were plotted\nin figure 44. NRTS curves for stability classes B and C are plotted for\nreference. The methyl iodide decrease of slope with distance - a de-\ncrease that was not explained simply by sampling inaccuracies or the\noccurrence of the peak concentrations at positions between samplers -\nshould be noted. Two additional curves, those labeled C and B\nLID\nLID'\nare shown to illustrate the effect upon the climatological curves\nif vertical tracer dispersion were confined within a layer 1400 m above\nground level. These two additional curves are discussed later.\nLateral standard deviations measured for methyl iodide were plotted\nversus downwind distance in figure 45. The NRTS curves for uranine dye\n(Yanskey et al. , 1966), labeled for stability classes B, C, and D, are\nincluded for reference. Statistically, the methyl iodide slope was the\nsame as the NRTS slopes of 0.85.\n65","5\n10\n10-6\nMethyl lodide\n10-7\nFigure 44. Methyl iodide measure-\nCLID\nments and NRTS climatological\nBLID\ncurves of relative axial concen-\ntrations versus downwind distances.\n8\n10\n100.0\n1000.0\n1.0\n10.0\nDistance (km)\nBC D\n4\n10\nTetroons\n3\nMethyl Iodide\n10\nB\nFigure 45. Lateral dispersion\nvalues from NRTS climatology,\nmethyl iodide measurements,\nand approximations from en-\nvelopes of tetroon\nI\n10\n100.0\n1000.0\ntrajectories.\n1.0\n10.0\nDistance (km)\nThe horizontal trajectories derived from the tower wind measurements\nare compared with the tetroon flights (fig. 46). . A tetroon trajectory is\nnot plotted for a release time of 1430 MST (the beginning of the methyl\niodide release). This flight ended about 60 sec after launch. Letter\nsymbols are located along the plotted trajectories at successive 15-min\nintervals, with times for windfield derived trajectories denoted by upper\ncase letters and tetroon times denoted by lower case letters. The tet-\nroon positions generally led the corresponding calculated trajectory posi-\ntions by a few kilometers (in agreement with pibal observed wind speeds\n66","Figure 46. Paired hori-\nlisted above each plot.\nletters) and windfield\nparticles. Trajectory\nzontal trajectories at\nletters) hypothetical\n15-min intervals for\ntetroons (lower case\nadvected (upper case\nbeginning times are\n(C) 1444 08/31/71\n(E) 1539 08/31/71\n(B) 1430 08/31/71\n08/31/71\n(A) 1303 08/31/71\n1507\n(D)","2 to 3 m sec-1 greater at tetroon flight altitudes). In every case, the\ntetroon trajectories deviated to the east of the windfield-determined tra-\njectories. The smallest lateral separation between a windfield-derived\ntrajectory and a tetroon trajectory occurred for the flight begun at 1507\nMST. This tetroon consistently flew at lower altitudes for more time than\nany other tetroon in this series.\nThe angular displacements between windfield and tetroon trajectories\nwere on the order of 5°. Pibal winds within the first 800 m above the\nsurface supported a wind direction turning consistent with this separation.\nThese same trajectories are shown in figures 47 and 48. In figure 47, all\ntetroon trajectories are plotted together. The dashed line is the wind-\nfield determined trajectory computed at the same time as tetroon release\nNo. 2. This dashed line seems to be a reasonable estimate of the extreme\nnorthwestward position of the tetroons. Open circles were plotted to\nnote the points of peak sampled methyl iodide on each arc. At 6 km, the\nwindfield derived trajectories (fig. 47) may be the best locaters of the\npeak concentration; at greater distances the tetroons were best (fig. 48).\nTo convert the trajectory data into a form similar to methyl iodide\nlateral dispersion, we arbitrarily assumed that the lateral width of the\nenvelope of the trajectories contained 95 percent of all the possible\ntra-\njectories that might have occurred in this period, or in other words, that\nthey represented +20 about the (undefined) mean. There is no justifica-\ntion for this assumption, although intuitively it seemed reasonable. A1-\nthough it rendered the absolute values of the plume spread suspect, the\n2 (WF)\nO 76 km - ARC D\n4\n5\nFigure 47. Composite plot of\ntetroon trajectories No. 3, 4,\n48 km - ARC C\nand 5. Dashed line locates\nwindfield derived trajectory\n3\nbeginning at 1430 MST. Loca-\ntions of peak-measured methyl\niodide concentrations are\nO 19 km ARC B\ndenoted by open circles.\nO\n6 km -\nARC A\n68","234 5\nI\nO 76km - ARC D\n48 km - ARC C\nFigure 48. Composite plot of\nwindfield derived trajector-\nies. Dashed line locates\ntrajectory for 1 hr before\nthe methyl iodide release.\no 19 km - ARC B\nLocations of peak measured\nmethyl iodide concentrations\n6 km - ARC A\nare the open circles.\nrate of spreading was unaffected by the particular choice of numerical O\nvalue. The line in figure 45 labeled tetroons is a plot of these envelope\nwidths after division by four. Since all of the tetroons were tracked\nshort of arc D at 76 km, the data point at 76 km was estimated from extrap-\nolations from the last trajectory end points. The vertical lines through\nthese points illustrate the maximum variability likely from uncertainties\nin defining the edges of the envelope of tetroon trajectories. When the\nvariability at each point was considered, the slope of these tetroon-\nderived o values did not significantly differ from either the NRTS curves\nor the methyl iodide measurements. The envelope-derived values are a fac-\ntor of 2 to 3 less than the methyl iodide measured lateral dispersion. In\nthe study of tetroon trajectories in the Los Angeles Basin, the trajectory\nenvelope width divided by 4 estimated the lateral dispersion determined by\nthe running mean variance statistic (Pack and Angell, 1963). This syste-\nmatic difference, always found to be less than a factor of 2, may explain\nsome of the differences in by values. When figures 47 and 48 were re-\nexamined, an interesting\nfeature was noted. When the envelopes of\nwindfield and tetroon trajectories were superimposed, the width of the com-\nbined envelopes was 1 12 to 2 times wider than the individual envelopes at\nthe longer distances. The average difference in angular bearing of the\ntrajectories (and their envelope means) was close to 5°. If lateral dis-\npersion was reestimated from the pooled envelopes of trajectories, values\nof O. approaching the methyl iodide measurements resulted. These syste-\nmatic differences in the near surface windfield trajectories and the cor-\nresponding tetroon trajectories suggested the effect of vertical wind\ndirection shear. Five sequential pibal observations, made at the release\n69","point and moving with the plume, substantiated a wind direction change of\nthis magnitude with height.\nLimiting o (the standard deviation of the vertical effluent concen-\nZ\ntration) to 80 percent of the depth of the mixing layer (Pasquill, 1962),\na maximum o of 1120 m resulted. When the effects of the vertical lid were\nincorporated to modify the rate change of normalized axial concentrations\nwith distance (Smith and Singer, 1966), a slower rate of dilution resulted.\nWhen the value of 1400 m for the lid was used to modify the curves in\nfigure 44 for NRTS stability classes B and C, the value of oz became a maxi-\nmum at 15 and 29 km, respectively. The effect of a lid appeared to reason-\nably account for the observed slower rate of dilution.\nAssuming a ground source (h = 0) and solving (6) for 'Z'\n(effective) = TUo Q Xp 1\n(12)\noz\n,\nwhere Q is rate of tracer release, U is the effective wind speed, Xp is the\nmeasured (methyl iodide) peak axial concentration, and o is the standard\ndeviation of methyl iodide in the lateral direction.\nFigure 49 shows\nthe calculated effectuve o values for methyl iodide. The line labeled\ntetroons represents the vatues of O. estimated from the envelope of lateral\nspreading of tetroons. Since the tetroon envelope derived values of o had\ny\nbeen shown to systematically underestimate lateral spreading, the cal\nculated effective o would obviously be an overestimate. This deviation\nZ\nfrom tracer-derived estimates should not be viewed as a failing of the\ntetroons; instead, it illustrates the crude (consistent) estimate possible\nby using only horizontal, unadjusted information. NRTS curves of O'z for\nstability classes B and C are plotted for comparison.\n10 4\nB\nC\nD\nFigure 49. Vertical dispersion\nvalues from the NRTS climatology,\nand the effective values deter-\nTetroon Estimated Limit\nI\nI\nmined from methyl iodide measure-\nAdiabatic Mixed\nLayer Estimated\nments. Estimates based on hori-\nLimit\n3\n10\nzontal tetroon trajectory envel-\nTetroons\nMethyl lodide\nopes used to define lateral dis-\npersion are labeled tetroons.\n2\n10\nB\nC\nD\n10\n1.0\n10.0\n100.0\n1000.0\nDistance (km)\n70","The dot-dashed line in figure 49 shows the maximum o value estimated\nZ\nfrom the depth of the adiabatically mixed layer. The\ndashed line\nlabeled tetroon estimated limit represents a maximum o equal to 80 per-\nZ\ncent of the height of the extreme tetroon vertical\noscillation. This\ntreatment was somewhat analogous to the arbitrary treatment of the lateral\nwidth of the horizontal trajectory envelopes. Since the tetroon envelope\nwidth divided by 4 underestimated the lateral dispersion, the \"tetroon\nenvelope\" effective O derived from (12) must have been overestimated by a\nfactor of 2 to 3. Likewise, use of 80 percent of the maximum attained\ntetroon height probably overestimated the effective height of limited ver-\ntical dispersion. The o determined from tetroon height oscillations was\n516 m. The mean effective o for methyl iodide tracer was near 650 m.\nSince the tetroon data are known to slightly underestimate vertical dis-\npersion, the tetroon O value was reasonable and as expected.\nThe various measurements from this test have shown that lateral tracer\nspreading behaved as expected. The relative axial concentration decreased\nmore slowly with distance than would be expected from an extrapolation of\nmeasurements at short distances. The effect of a lid, which inhibited\nvertical dispersion, seemed to explain this slowed rate of dilution. The\ntracer centerline trajectory was well described by both the windfield de-\nrived trajectories and the tetroon trajectories. At the longer distances\nand in the presence of vertical shear, the tetroon trajectories were the\nbest indication of transport. The lateral tracer dispersion can probably\nbe well approximated by the running mean variance statistic; less detailed\ninformation such as the trajectory envelope width should be used only with\nconsiderable caution. The standard deviation of tetroon height oscilla-\ntions provided a reasonable estimate of the mean effective o for the\nZ\nmethyl iodide tracer.\n11. DIAGNOSTIC APPLICATIONS OF WIND SPEED AND COMPONENT SPECTRA\nThe comparison of the energy spectra of horizontal wind speed with\nthe summed spectra of the horizontal wind components has been shown to pro-\nvide some interesting insight into the fluctuating nature of the wind at a\nsingle location (NOAA Tech. Memo. ERL ARL-32, 1971). It was demonstrated,\nwith a model wind composed of a simple harmonic, that the energy distri-\nbutions for speed and component velocity can be drastically different.\nThe energy distribution for the velocity had all the energy appearting at\nthe frequency of the oscillation in the velocity. The energy distribution\nfor the speed had less than 20 percent of the energy appearing at twice\nthe frequency and the rest appearing as the energy of the mean flow (zero\nfrequency). The reason for the difference was that the speed in the model\ncould not reflect the sign change in the velocity due to direction rever-\nsal. Applying this information to the energy spectra of real data, one\nwould surmise that, if the speed energy in a given frequency band is much\nless than the combined component energy, the wind fluctuations in that\nfrequency band would be caused by direction reversals. If there is not\nmuch difference in the speed energy and combined component energy in a\n71","given frequency band, then the wind fluctuations would be primarily due\nto fluctuations in the speed alone.\nThe single harmonic approximation to the wind is an extremely simpli-\nfied version of what occurs in reality, as may be observed from the spec-\ntral distributions of atmospheric kinetic energy. The variations in the\nmotion of the air flowing past a point in the free atmosphere are influenced\nby phenomena of several scales. For transport considerations, the primary\ninfluences seem to be the synoptic scale weather patterns, and local scale\ntopographic influences (including land-water proximity). The combination\nof these influences can cause a time series of wind data from a given\nlocation to be quite complex. In figures 50 through 53, time series of\nwind components and speed are plotted for July and February 1969, for\ntwo levels on the CFA tower at the NRTS. For comparison, the data shown\nare from the data set used to obtain the energy spectra for 1969 (NOAA\nTech. Memo. ERL ARL-32, 1971). The coordinate axes have been rotated 49°\nclockwise to provide a clear look at the variation along the principal\naxis of oscillation (V component). Figure 50 shows that at 76 m during\nFebruary, the velocity variation involves a large number of NE-SW - direction\nreversals with periods around 2 to 4 days. During July, figure 52 shows\nthe prominent velocity variation is a diurnal reversal of the wind. These\nsignificant wind reversals cause the speed traces in both of these cases\nto appear quite different from the velocity traces. The direction\n(A) TIME SERIES OF CFA 76 M WINDS FOR FEB 1969, RTE=49\n30\n20\n10\n0\n-10\n30\n20\n10\n0\n-10\n-20\n-30\n30\n20\n10\n0\n-10\n-20\n-30\n16\n17\n18\n19\n20\n21\n22\n23\n24\n25\n26\n27\n28\n0\n1\n2\n3\n4\n6\n7\n8\n9\n10\n11\n12\n13\n14\n15\nTIME (DAYS)\nFigure 50. Plots of time variation during February 1969, of\nwind components and wind speed for the 76 m level of the\ntower at the Central Facilities Area (CFA) of the NRTS.\nThe 49° clockwise rotation of the coordinate areas causes\nthe V component to represent a NE or SW wind.\n72","(B) TIME SERIES OF CFA 6 M WINDS FOR FEB 1969. RTE=49\n30\n20\n10\nS\n0\n-10\n30\n20\n10\n0\n-10\n-20\n-30\n30\n20\n10\n0\n-10\n-20\n-30\n8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28\n0\nTIME (DAYS)\nFigure 51. Same as fig. 50 except for 6-m level.\n(A) TIME SERIES OF CFA 76 M WINDS FOR JULY 1969, RTE=49\n30\n20\n10\n0\n-10\n30\n20\n10\n0\n-10\n-20\n-30\n30\n20\n10\n0\n-10\n-20\n-30\n0\n3\n8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30\n31\n.\nTIME (DAYS)\nFigure 52. Same as fig. 50 for July 1969.\n73","(B) TIME SERIES OF CFA 6 M WINDS FOR JULY 1969, RTE=49\n30\n20\nSPO\n10\nS\n0\n-10\n30\n20\n10\n0\n-10\n-20\n-30\n30\n20\n10\n0\n-10\n-20\n-30\n10\n11\n12\n13\n14\n15\n16\n17\n18\n19\n20\n21\n22\n23\n24\n25\n26\n27\n28\n29\n30\n31\n0\n1\n2\n3\n8\n9\nTIME (DAYS)\nFigure 53. Same as fig. 51 for July 1969.\nreversals of many frequencies compared with many other smaller amplitude\noscillations cause the redistribution of energy in the speed spectra to\nbe more subtle and complex than for the modeled wind containing a re-\nversal at a single harmonic. However, the predominant effect for those\nfrequencies involving direction reversals is a significant reduction of\nenergy in the speed spectrum at those frequencies. Apparently, for a\ngiven frequency band, one can safely attribute the loss of energy in the\nspeed spectra to direction reversals in the wind.\nComparing the spectra at 6 m showed that the speed energy was about\nthe same as the component energy for the frequency band containing the\ndiurnal cycle. The reason for this may be seen by comparing the plots\nin figures 52 and 53. The velocity fluctuations at 6 m involve a much\nless significant amount of direction reversal than observed at 76 m.\nThis causes the 6 m speed and component trace to resemble the diurnal var-\niations closer. Examining these time series plots was to substantiate the\nfindings previously deduced from the spectral comparisons. Even though\nthe spectral comparisons yield only qualitative information, they provide\na quick and inexpensive method to investigate the importance of local\nterrain features on boundary layer flow.\nFor the data on the CFA tower, the spectral comparisons indicated a\nstrong possibility for an unusual amount of low-level shear. When the\nlocal terrain around the tower is considered, the cause of the phenomenon\nbegins to emerge. A relief map of the upper Snake River Plain containing\nthe location of the CFA tower is shown in figure 54. Note that the general\n74","4\nb\nFigure 54. Relief map of the upper Snake River Plain in SE Idaho.\nThe values shown on the few contour lines over the plain are\nin hundreds of feet. The stippled area within and adjacent\nto the grid indicates the area below 5000 ft MSL. The tics\nalong the border of the grid indicate the grid point separa-\ntion, 5.33 mi. (a) CFA 76-m tower location, (b) Grid III\n61-m tower location.\n75","slope near the tower is from SW to NE. The magnitude of this local slope\nis apparently great enough to significantly oppose the general drainage\nflow from the NE. A dramatic example of this is shown in figure 29 in\nwhich an early morning smoke release from different levels on a tower\nabout 6 km north of the CFA tower was photographed. The smoke from the\ntop level (61 m) is traveling SW at about 8 m sec-1, while the smoke\nfrom the bottom is traveling ENE at about 2 m sec-1. The results of the\nspectral comparison indicate that this is not a rare phenomenon.\nAn opportunity arose to apply the diagnostic spectral techniques to\nsome data from a 61-m tower on the Arnold Engineering Development Reserva-\ntion in south central Tennessee. The terrain variation near the tower is\nsimilar to that in central Oklahoma with streambed erosion causing a mod-\nerate amount of small-scale roughness. On a scale of tens of kilometers,\nthe terrain is quite flat except for an abrupt height change at the begin-\nning of the mountain range about 16 km to the east. To illustrate the\nproximity of the tower to the mountain range, the 1100 ft (330 m) and 1900 ft\n(570 m) contours are shown in figure 55. In many areas, this 800-ft (240\nm)\nchange in elevation occurs in less than 2 km, but then the high ground\nabruptly levels off to the east. Few heights are above 2,000 ft (600 m).\nIf we were trying to determine what type of flow might be induced by\nthe terrain under conditions of weak synoptic pressure gardient, the ques-\ntion of whether drainage flow from the mountains would reach the tower\nlocation would be difficult to answer. The abruptness and brevity of the\nslope and its separation from the tower by several miles of relatively\nflat ground would raise doubt that the wind measured at the tower would\nreflect a diurnal effect caused by the mountains to the east.\nThe wind data from the 61 - and 10-m levels taken during 1964 were\nused in a spectral comparison test and are shown in figures 56 and 57,\nRelatively Flat\nTerrain\nFigure 55. Relative locations of\nthe Arnold Engineering Develop-\nment Center (AEDC) tower with\nrespect to relatively flat\nAEOC\nTullahoma\nTower\nground and mountainous terrain.\nThe dashed line shows the orien-\ntation of the principal axis of\noscillation of diurnal cycle in\nMountainous\nthe wind at 61 M.\nTerrain\n0\n5\n10\nScale (miles)\n76","PERIOD (HR)\n0\n3\n10 ²\n10\n10\n10\n20\n61M SPECTRA AFA, TENNESSE\nU+V\nSPD\nKm\nKe\nKt\nU+V\n1.3\n20.8\n22.0\nSPD\n16.5\n5.6\n22.0\n10\n0\n10\n10\n10\n10°\nFREQUENCY (CYCLES HR-1,\nFigure 56. Energy spectra for wind speed and combined spectra\nof the u and V components of the wind at 61 m on the AEDC\ntower for 8192 hours beginning December 1, 1964.\nPERIOD (HR)\n4\n3\n2\n10 1\n0\n10\n10\n10\n20\n10M SPECTRA AFA, TENNESSE\nU+V\nSPD\nKm\nKe\nKt\nU+V\n0.3\n7.4\n7.6\nSPD\n4.8\n2.9\n7.6\n10\n0\n10°\n10\n10\n10\n10\nFREQUENCY (CYCLES HR-1,\nFigure 57.\nSame as fig. 56 except at 10 m.\nrespectively. In comparing these results with those for the NRTS, we find\nthe total energy at the Tennessee tower to be little more than half that\nat the NRTS. According to a logarithmic extrapolation, the 15-m\nheight\ndiscrepancy accounts for only a small fraction of the difference. The\nmean wind speed for the Tennessee data was 4. 1 m sec-¹ and for the Idaho\ndata, it was 5.0 m sec-1. The spectral distribution of component energy\nfor the 61-n level in the Tennessee tower shows the contribution to the\neddy energy from synoptic disturbances coming from a narrower band of fre-\nquencies (periods from 2 to 3 days) than at the NRTS (1 5 to 14 days).\nThis difference could reflect the effect of the mountains as cyclonic\nstorms pass through the Northwest.\n77","The most striking difference in the Idaho and Tennessee data at the\n76- and 61-m levels is the relative size of the energy contribution from\nthe diurnal cycle. The diurnal cycle in the wind variation measured at\nthe tower in Tennessee is significant but much less so than at the Idaho\ntower. Since no indication of a peak at the diurnal frequency appears in\nthe speed spectra, the diurnal variation in the speed can be attributed\nto a wind reversal, as was the case for the NRTS data. The spectral com-\nparison for 10-m level of the Tennessee data indicates that the diurnal\nvariation is due primarily to a speed fluctuation rather than a direction\nreversal.\nIn contrast to the 6-m Idaho data, the energy contribution of the\nspeed spectrum at the 10-m level in the Tennessee data is significantly\ngreater than the energy contribution of the combined component spectra.\nThis result is similar to one obtained for Caribou, Maine (Oort and\nTaylor, 1969). One difference is that, in a line-by-line investigation\nof the spectra, they found the diurnal spike in the speed spectra to be\ntwo orders of magnitude greater than in the component spectra. A com-\nparison of this type for the Tennessee data shows the spike in the speed\nspectra to be only a factor of 3 greater than in the component spectra.\nAs a test of the significance of the spikes in the component spectra, we\nrecomputed them using the same time-varying direction but a constant value\nin place of the time-varying speed. The result showed no significant de-\ncrease in the relative magnitude of the spike in the component spectra.\nThis would indicate that, at least in this case, the extra energy in the\nspeed spectra at the diurnal cycle is a reflection \"beating\" by direction\nreversals at other frequencies.\nThe directional nature of the wind variation was examined by recal-\nculating and plotting the component spectra at 61 m while rotating in 10°\nincrements the coordinate axes clockwise through 80°. The spectra for\nthe N-S orientation and the 40° rotation are shown in figures 58 and 59.\nFor the N-S orientation, the total energy for the V component (along the\nordinate) is over twice that for the u component (along the abscissa).\nThe plot clearly shows that the variation in the N-S component of the\nwind is the major contributor to the eddy energy over the range of the\nsynoptic scale disturbances, but not at the diurnal cycle. As the axes\nwere rotated clockwise, the difference in contribution at the diurnal\ncycle was a maximum at 40°, with the u component being significantly\nlarger. This would indicate a diurnal oscillation in a SE-NW direction.\nIt is interesting to note the relative balance between the energy contri-\nbutions of the components through the range of frequencies for the synoptic\ndisturbances. This contrasts with the situation for the Idaho data, in\nwhich the principal axis of oscillation was the same for both the synop-\ntic disturbances and diurnal oscillations because of the strong channel-\ning effect of the mountains on either side of the Upper Snake River Basin.\nIn relating the spectral results for the Tennessee data to the topo-\ngraphy, we find that N-S orientation of the principal axis of oscillation\nfor the synoptic range of frequencies is probably caused by some deflection\n78","PERIOD (HR)\n0\n3\n10 ²\n10\n10\n10\n20\n61M SPECTRA AFA, TENNESSE\nU\nV\nKm\nKe\nKt\nRTE 00\nU\n0.1\n6.5\n6.6\nV\n1.2\n14.3\n15.4\n10\n0\n10°\n10-3\n2\n10\n10\n10\nFREQUENCY (CYCLES HR - 1,\nFigure 58. Energy spectra for the u and V components\nwith V being the N-S component.\nPERIOD (HR)\n4\n10 3\n2\n10 1\n10 0\n10\n10\n20\n61M SPECTRA AFA, TENNESSE\nU\nV\nKm\nKe\nKt\nRTE\n40\nU\n0.2\n9.7\n10.0\nV\n1.0\n11.0\n12.1\n10\n0\n10-3\n102\n10\n10\n10°\nFREQUENCY (CYCLES HR 1,\nFigure 59. Energy spectra for the u and V components\nwith V being the NE-SW component.\nof the flow by the mountain range to the east. The NW-SE orientation\nof the diurnal oscillation indicates a mountain-valley effect, at least\nat the 61-r level, due to the valley SW of the tower. The effect seems to\nbe nullified at 10 m, since only a speed fluctuation occurs at this level.\nThis is probably due to the relatively flat terrain between the tower and\nvalley as well as a large reservoir about 5 km south of the tower.\nSeasonal spectral comparisons for 1964 show about what to expect.\nThe diurnal oscillation varies from undiscernable in the winter to\nthe\nmajor contribution from a single frequency band in the summer. The total\nvariance is a minimum in summer (10.6 m²sec-2 and a maximum in winter\n(31.3 m² sec-2 This difference can be attributed to the large variation\nin the contributions from the synoptic scale frequencies.\n79","To summarize these results in terms of transport, we could say that\na moderate preference for transport to the north or south exists during\nthe fall, winter, and spring with the predominant periods for wind re-\nversals being 4 to 8 days during the spring and fall, but 2 to 4 days dur-\ning the winter. During the warm months, transport for a stack release\nwould be strongly affected by the diurnal circulation from the valley to\nthe SE. This demonstrates again that, for the meteorological aspects of\nreactor safety analyses, diagnostic spectral techniques provide valuable\ninsight concerning the influence that local topographic features have on\nthe wind at a given location. Such information pertaining to the oscil-\nlatory nature of the wind would be impossible to obtain with a standard\nwind rose analysis.\n12. FORECASTING AND WEATHER WARNING SERVICE\nThe number of specific NRTS subcontractor and contractor requests\ncan be estimated as follows. About six subcontractors ran projects in\nwhich daily or twice-daily requests for weather information were made for\nperiods of 2 to 10 weeks. This represents about 200 to 250 forecast re-\nquests. There were 13 weather warnings (excepting December and January)\ngiven to AEC Warnings and Communications for general broadcast at the\nNRTS. December 1971 and January 1972 were the stormiest of record, and\n29 individual weather warnings were issued in the 2 months. A descrip-\ntion of the unusual wind storm of January 11 and 12, in which the NRTS\ntransportation buses were stranded at the NRTS overnight, follows. Re-\nquests for weather advisories and forecasts occurred between two and\nthree times daily as a minimum, and upwards of 30 requests daily during\nthe unusual weather. Admittedly, a number of these requests were from\nindividuals satisfying their own curiosity; however, a large percentage\nwere from administrators of NRTS functions such as plant operations,\ntransportation, repair services, etc.\nCertain teletype and facsimile data - normally received to provide\nthe weather warning and advisory services - are also filed for possible\nfuture reference in research on transport.\nThe storm of January 11 and 12, 1972, involved about 36 man-hours\nof closely watching the available data, issuing warnings, personally con-\nsulting with management, and cooperating with Warnings and Communications.\nFrom late morning of Tuesday, January 11, to the predawn hours of\nWednesday, January 12, we recorded the strongest winds over snow-covered\nground ever observed at the NRTS. These record-breaking winds were due\nto a possibly record-breaking surface pressure gradient between the high\npressure in the intermountain region and the low perssure on the east\nside of the Rocky Mountains. At 1700 MST of January 11, the low was in\ncentral Montana with a central pressure of only 976 mb, while Pocatello,\nIdaho, was still at 1006 mb - a 30 mb difference. Most of the difference\nwas concentrated within about 500 km. Table 7 shows the average hourly\n80","Table 7. Hourly Average Winds and Peak Gusts Recorded During\nthe Hour ending at the Time Indicated (direction-speed-peak\ngust).\nCFA\nGrid III\nTREAT\nTime (MST)\n20' Level\n200' Level\n50' Level\nJanuary 11\n0100\n236-21-30*\n235-32-40\n225-18-31\n0200\n252-14-23\n240-32-40\n225-20-30\n0300\n252-19-38\n250-22-33\n225-25-38\n0400\n234-30-41\n260-18-34\n225-28-42\n0500\n230-31-43\n235-35-45\n228-31-42\n0600\n231-27-40\n235-44-53\n235-38-52\n0700\n238-31-46\n237-49-59\n235-38-52\n0800\n236-33-47\n245-53-67\n235-38-52\n0900\n237-36-53\n245-59-68\n238-40-54\n1000\n238-38-55\n240-59-70\n237-44-56\n1100\n237-37-51\n240-52-68\n237-42-56\n1200\n234-34-49\n235-49-60\n240-43-64\n1300\n234-36-60\n232-48-60\n235-51-67\n1400\n234-40-61\n238-52-66\n235-53-68\n1500\n237-44-61\n240-61-72\n235-46-61\n1600\n237-44-63\n240-64-79\n240-47-63\n1700\n240-43-61\n242-64-77\n240-49-68\n1800\n240-44-61\n243-65-76\n241-47-65\n1900\n244-44-60\n245-63-75\n245-50-69\n2000\n245-46-70\n245-65-79\n245-51-71\n2100\n244-43-70\n245-62-76\n245-48-66\n2200\n243-44-65\n245-64-77\n245-47-63\n2300\n243-48-64\n245-63-79\n245-46-61\n2400\n243-42-63\n245-56-67\n245-45-61\nJanuary 12\n0100\n241-38-57\n245-54-65\n245-44-58\n0200\n240-36-54\n243-50-67\n245-42-56\n0300\n244-37-53\n245-52-62\n245-40-55\n0400\n246-35-51\n248-41-57\n252-37-49\n0500\n261-27-43\n265-30-47\n255-30-43\n0600\n270-18-30\n300-13-30\n320-10-30\n0700\n351-10-21\n335-16-34\n350-5-10\n* Direction in degrees, speed in knots.\n81","winds and peak gusts for a 30-hr period at three locations starting the\nmorning of January 11 and ending the morning of the 12th. Table 8 shows\nthe altimeter setting differences at selected hours. A difference of\n0.10 inches of Hg between Pocatello and Idaho Falls is unusual, and usually\nproduces average wind speeds of 30 to 35 mph.\nTable 8. Altimeter Setting Differences in Inches of Mercury\nPIH-IDA\nDate\nIDA-DLN\nHour\nJan\nMST\n10\n2300\n0.03\n0.16\n11\n0500\n0.09\n0.25\n11\n0800\n0.11\n0.30\n11\n1100\n0.11\n0.35\n11\n1400\n0.11\n0.32\n11\n1700\n0.20\n0.25\n11\n2000\n0.18\n0.19\n11\n2300\n0.16\n0.08\n12\n0200\n0.09\n0.01\n12\n0500\n0.10\n-0.06\n12\n0800\n-0.02\n-0.02\nPIH - Pocatello, Idaho\nIDA - Idaho Falls, Idaho\nDLN - Dillon, Montana\n13. ACKNOWLEDGMENTS\nSections 1 and 2 were joint research efforts of several Air Resources\nLaboratories with the sponsorship of both NOAA and the AEC Division of\nReactor Development and Technology. Section 5 is work sponsored by the\nAEC Directorate of Licensing.\nThe Oklahoma field program was carried out by ARL personnel from\nSilver Spring, Maryland, and the National Reactor Testing Station (NRTS),\nIdaho Falls, Idaho. Support was provided by the National Severe Storms\nLaboratory (NSSL), Norman, Oklahoma, which operated the slow-response tower\nsensor system and provided assistance in numerous other ways, and by the 6th\nWeather Squadron (Mobile), Air Weather Service, U.S. Air Force, which operated\nthe ibal-radiosonde network.\n82","14. REFERENCES\nAngell, J. K. , D. H. Pack, C. R. Dickson, and W. H. Hoecker (1971), Urban\ninfluence on nighttime air flow estimated from tetroon flights, J.\nAppl. Meteorol. 10:194-204.\nBellamy, John C. (1945), The use of pressure altitude and altimeter cor-\nrections in meteorology, J. Meteorol. 2(1):1-79.\nBonner, W. D., and J. Paegle (1970), Diurnal variations in the boundary\nlayer winds over the south central United States in summer, Monthly\nWeather Rev. 98(10):735-744.\nBriggs, G. A. (1969), Plume rise, U.S. Atomic Energy Commission, Division\nof Technical Information, TID-25075.\nCamp, D. W., R. E. Turner, and L. P. Gilchrist (1970), Response tests of\ncup, vane, and propeller wind sensors, J. Geophys. Res. 75:5265-5270.\nDingle, A. Nelson, and Charles Young (1965), Computer applications in the\natmospheric sciences, Univ. of Michigan, Ann Arbor, Mich., 214-221.\nDrinkrow, R. (1972), A solution to the paired Gill-anemometer response\nfunction, J. Appl. Meteorol. II:76-80.\nGill, G. C. , L. E. Olsson, J. Sela, and M. Suda (1967), Accuracy of wind\nmeasurements on towers or stacks, Bull. Am. Meteorol. Soc. 48:665-674.\nHeffter, J. L. (1965), The variation of horizontal diffusion parameters\nwith time for travel periods of one hour or longer, J. Appl. Meteorol.\n4(1):153-156.\nHolmes, R. M. , G. C. Gill, and H. W. Carson (1964), A propeller-type ver-\ntical anemometer, J. Appl. Meteorol. 3:802-804.\nKaimal, J. C., and D. A. Haugen (1969), Some errors in the measurement of\nReynolds stress, J. Appl. Meteorol. 8:460-462.\nNOAA Tech. Memo (1970), Atmospheric transport and diffusion in the plane-\ntary boundary layer, Air Resources Laboratories Semi-Annual Research\nProgram Review, Jan-July 1970, ERL-ARL-28, 47 pp.\nNOAA Tech. Memo (1971), Atmospheric transport and diffusion in the plane-\ntary boundary layer, Air Resources Laboratories Annual Research Pro-\ngram Review, July 1970-June 1971, ERLTM-ARL-32, 71 pp.\nOort, A. H. , and A. Taylor (1969), On the kenetic energy spectrum near\nthe ground, Monthly Weather Rev., 97:632-636.\nPack, D. H. , and J. K. Angell (1963), A preliminary study of air trajec-\ntories in the Los Angeles Basin as derived from tetroon flights,\nMonthly Weather Rev. 01(10-11):583-604.\nPasquill, F. (1962), Atmospheric Diffusion (D. Van Nostrand Co., Ltd., ,\nLondon).\n83","Slade, D. H., , ed. (1968), Meteorology and atomic energy 1968, TID-24190,\nClearinghouse for Federal Scientific and Technical Information, Natl.\nBureau of Stands., U.S. Dept. of Commerce, Springfield, Va.\nSmith, M. E., , and I. A. Singer (1966), An improved method of estimating\nconcentrations and related phenomena from a point source emission,\nJ. Appl. Meteorol. 5(5):631-639.\nStart, G. E. (1970), Comparative diffusion and deposition of uranine dye,\nmolecular iodine gas, and methyl iodide gas, Proc. 5th Annual Mid-\nyear Topical Symp. on Health Physics of Nuclear Facility Siting,\nHealth Physics Soc. Nov.\nTeweles, S. , and H. Wobus (1954), , Verification of prognostic charts, Bull.\nAm. Meteorol. Soc. 35:455-463.\nWendell, L. L. (1970), A preliminary examination of mesoscale wind fields\nand transport determined from a network of towers, NOAA Tech. Memo.\nERLTM-ARL 25, Air Resources Laboratories, Silver Spring, Md.\nWendell, L. L. (1972), Mesoscale windfields and transport estimated deter-\nmined from a network of wind towers, Monthly Weather Rev. 100(7):565-\n578.\nYanskey, George R., , Earl H. Markee, Jr., , and Alden P. Richter (1966),\nClimatography of the national reactor testing station, IDO-12048,\nAEC Opeartions Office, Idaho Falls, Idaho.\n84\nUSCOMM -"]}