{"Bibliographic":{"Title":"Numerical weather prediction activities report.","Authors":"","Publication date":"1975","Publisher":""},"Administrative":{"Date created":"08-17-2023","Language":"English","Rights":"CC 0","Size":"0000097804"},"Pages":["QC\n875\nUNIVERSITY\nOF\nCOMMUNITY\nUNITED STATES DEPARTMENT OF COMMERCE\n.U54\n1975\nNational Oceanic and Atmospheric Administration\nAvenue\nWATED\nNational Weather Service\nSTATES OF\nNUMERICAL WEATHER PREDICTION\nACTIVITIES - 1975","UNITED STATES DEPARTMENT OF COMMERCE\nNATIONAL OCEANIC AND ATMOSPHERIC ADMINISTRATION\nNATIONAL WEATHER SERVICE\nNUMERICAL WEATHER PREDICTION ACTIVITIES REPORT\n1975\nCONTENTS\nPART\nSUMMARY OF MAJOR HIGHLIGHTS OF RESEARCH APPLICATIONS\nI\n1\nAND OPERATIONAL CHANGES\nRESEARCH AND DEVELOPMENT IN ANALYSIS-PREDICTION SYSTEMS\nII\n(Development Division, National Meteorological Center)\n3\nTECHNIQUES, DEVELOPMENT AND APPLICATIONS OF NEW PRODUCTS\nIII\n(Systems Development Office, Techniques Development Lab.) 24\n31\nANALYSIS-FORECAST SYSTEMS IN OPERATIONAL USE IN 1975\nIV\n33\nPLANS FOR FUTURE OPERATIONAL SYSTEMS\nV\n34\nFORECAST VERIFICATIONS\nVI\n40\nPUBLICATIONS\nVII","Q C\n875\nU54\nUNITED STATES OF AMERICA\n1975\nNational Weather Service\nNational Oceanic and Atmospheric Administration\nPART I\nSUMMARY OF MAJOR HIGHLIGHTS OF RESEARCH APPLICATIONS AND\nOPERATIONAL CHANGES\n1.1\nThis report summarizes the 1975 activities in numerical weather\nprediction of the National Meteorological Center and, for the first\ntime, the activities of the Techniques Development Laboratory of the\nNational Weather Service of the USA.\n1.2\nThis year was the twentieth anniversary of the installation\nand use of the first electronic computer in the USA for real-time\nnumerical weather prediction. A few who were in the original staff\ntook note of this milestone and the progress which has been made in\noperational numerical weather prediction during these past 20\nyears. The basic forecasting function of the National Weather Service\nhas been significantly influenced over the years by this forecasting\nmethod, which originated in this early effort.\n1.3\nThe following major events took place during 1975:\n(a)\nThe primary computer facility was upgraded by the acquisition\nof a third IBM 360/195 computer late in the year. The major organiza-\ntional components of the National Oceanic and Atmospheric Administration,\nprimarily the National Weather Service and the National Environmental\nSatellite Service, used the dual IBM 360/195 system for their\noperational requirements throughout most of the year. The processing\nof the meteorological data was done on the NMC system which consisted\nof three IBM 360/40's and one IBM 360/30.\n(b)\nOn May 21, 1975, the forecast period of the limited-area fine-\nmesh model (LFM) was extended from 24 hours to 36 hours for routine NWS\nuse. The forecast superiority of this prediction system over the\noperational hemispheric system (6L PE) was especially apparent during\n1975. On December 17, 1975, tests were initiated for the possible\nextension of the LFM runs to 48 hours.\n(c)\nOn November 4, 1975, the plotting of the North America surface\ncharts was automated. The conventional plotting model was used, and\nthe charts produced at eight observing times each day. Automation of\nthe plotting on the Northern Hemisphere surface charts is planned for\n1976.","-2-\nAutomation of many of the charts for facsimile transmission was\n(d)\ncompleted. Analyses and forecasts produced on the IBM 360/195's are\nnow transferred directly onto the facsimile circuits without human\nintervention. The relatively few maps which are manually prepared in\nthe World Weather Building are automatically digitized and transmitted\ninto the computer system at Suitland for automatic transmission to the\nfield forecaster.\nA ten-layer limited-area model designed for short-range\n(e)\nhurricane forecasting was tested in near-real-time during the 1975\nhurricane season.\nThe first phase of the Data Systems Test (DST) was completed on\n(f)\nthe August and September data. The second phase will be conducted\nduring January and February 1976. The DST is a USA test to ensure an\nearly consideration of many of the problems to be confronted in the\nFirst GARP Global Experiment (FGGE). .\nA physical oceanography group has been established within NMC.\n(g)\nProject areas are: 1) numerical modeling of coastal circulation;\n2) specification of meteorological forcing for coastal models;\n3) objective analysis of sea surface temperature; and 4) wave refraction.","-3-\nPART II\nRESEARCH AND DEVELOPMENT IN ANALYSIS-PREDICTION SYSTEMS\n(Development Division, National Meteorological Center)\n2.1\nGlobal Modeling Branch\n2.1.1\nChanges to FINAL 8-1ayer Global Model (8L GLOBAL (2.5°)\nThe major changes to the 8L GLOBAL (2.5°) run at FINAL time\nfollow:\n(a)\n50-mb analyzed winds, rather than calm winds, are used as\ninitial conditions in the \"thetasphere\"- the computational layer at\nthe top of the model. The change was introduced on January 16, 1975.\n(Stackpole)\n(b)\nThe sigma-to-pressure coordinate transformation was\nsubstantially revised and the changes incorporated into the code May 28,\n1975.\n(Stackpole)\n(c)\nEfforts by F. Shuman have resulted in a changed design for the\ntriangular tendency smoother. It is applied to the tendencies on the\nlongitude-latitude grid and assures that the Courant-Frederichs-Lewy\nlinear instability conditions'are not violated. A slight departure\nfrom a precisely triangular shape has eliminated the possibility of a\nnegative response for particular wave lengths. A theoretical method\nof specifying the quasi-triangle base-widths has been developed also.\nThis resultant width is only slightly less than that which was\npreviously computed experimentally. It is the smallest that can be\nused and satisfies the CFL stability requirements. This small width\nminimizes truncation error. This technique was introduced on May 28,\n1975.\n(Stackpole)\n(d)\nA full-latitude smoother, developed by F. Shuman and\nspecifically designed for latitude-longitude grids, has been applied\nto the output fields of the 8L GLOBAL. This smoother maintains the\nsame east-west resolution, as the meridians converge, as is maintained\non the north-south direction. The smoother is full latitude [all-the-\nway-around] with a very sharp drop in response at the cutoff frequency.\nThe smoothing is used for the output preparation of forecast maps.\nThe technique was introduced on May 28, 1975.\n(Stackpole)","-4-\nAn error was corrected in the calculation of pressure heights\n(e)\nunder the model terrain; the error, which did not affect the 1000-mb\nheights, caused the 850- and 700-mb heights to be too low by 30 or 40\nmeters. Since the 1000-mb heights were computed correctly, 850-1000 mb\nthicknesses, \"under\" mountains, were distorted. The correction was\napplied at those geographical points where the ground elevation is\nabove the height of the 850- and 700-mb levels. This correction was\n(Stackpole)\nintroduced on August 12, 1975.\nThe Brown-Phillips energy-conserving form of the hydrostatic\n(f)\nequation replaced the old method of calculating heights at sigma levels\nand of T in the layers. This form of the equation, plus the related K\nspecification of the value of the Exner function, TT = (p/1000)k , at the\n\"center\" of the sigma layers, was coded by William Collins. Tests\nindicate a slightly better behavior of the total kinetic and potential\nenergies. These modifications were incorporated into the operational\n8-1ayer code on November 20, 1975. NMC Office Notes 92 (Brown) and 104\n(Phillips) describe this form of the hydrostatic equation.\n(Stackpole and Collins)\nThe parameterization of subgrid-scale convective precipitation\n(g)\nwas incorporated into the model. This method of convective parameterization\ncalculates the latent heat and precipitation contribution of subgrid-scale\ndeep convection to the large-scale system. Eight-layer forecasts of\nconditionally unstable areas with favorable upward motion initiate the\nconvective process. The subgrid storm extends vertically throughout the\ntroposphere; the moist adiabat (from the lifted condensation level\ndefined by air-parcel characteristics of the lowest model layer) pre-\nscribe temperature and moisture distributions. The storm wind has no\nvertical shear and is defined by a vertical mix of the large-scale\nprofile. Storm upward motion is estimated from the existing large-scale\nconvergence field; the extent to which the large-scale conditions depart\nfrom the ideal (for convective activity) determines the fractional area\nof the storm. This change was incorporated on December 10, 1975. (Hirano)\nModifications to the large-scale precipitation method relax the\n(h)\nformer saturation-evaporation criterion. The former method required\nsufficient evaporation of rain as it fell through an unsaturated layer\nto bring that layer to saturation. The new model has a vertically\nvarying condensation-precipitation criterion (-100% relative humidity)\nsuch that any moisture forecast in excess of the criterion in a model\nlayer falls from that layer as rain. If this rain falls into the next\nlower layer and that layer is less than saturated with respect to a\nseparate saturation-evaporation criterion, then enough rain will evaporate\nto bring that lower layer up to this separate saturation level. This\n(Hirano)\nmodification was introduced on December 10, 1975.","-5-\n(i)\nAs a side effect of changes (f) and (g) above, the \"moist con-\nvective adjustment\" was deleted from the model on December 10, 1975.\n(j)\nA difficulty with forecasts of longer than 2 days or so in the\nvicinity of the poles was alleviated by specifying (rather than fore-\ncasting) the polar values as the circumpolar average of the quantities\nat the adjacent row. This change was introduced on December 10, 1975.\n(Stackpole)\n2.1.2 Testing of 2.0° Hemispheric Version of 8-1ayer Model\nUsing the physical-numerical alterations detailed in 2.1.1, a\nweekly series of forecasts from the 8L HEM (2°) were implemented; these\nforecasts were extended to 84 hours. Depletions of mass in the top\nlayer of the model at some grid points resulted in model failure. This\nproblem was remedied by the use of a new upper boundary condition: a\nspace and time constant pressure of 50 mb. (It should be noted that\nthe 8L GLOBAL (2.5°) model, at FINAL time, still retains the layer of\nconstant potential temperature as its upper boundary condition.) These\ntests also included a divergence damper, applied to the stratospheric\nwinds, to control noise. (See 2.1.6.) The model uses a 7.5-minute\ntimestep.\nEach forecast was compared directly with its operational 6L PE\ncounterpart. The Global Modeling Branch compiled standard statistical\nverification data for the 8L HEM (2°) as did the Data Assimilation\nBranch for the 6L PE, and the two sets of data were then compared. In\naddition a Forecast Division jury evaluated the 24-, 48-, and 84-hour\nforecasts subjectively. R. Hirano undertook the verification of the\nprecipitation forecasts.\nFour forecast cycles were completed by year's end and the program\nis projected to continue through March 1976.\n2.1.3\nRevisions of the Sigma-to-Pressure Coordinate Transformation\nin the 8L GLOBAL (20) Model\nIn constructing mandatory-level parameters from the sigma-\ncoordinate system, two methods are used: mandatory-level winds and\ntemperatures are obtained by linear interpolations from sigma layers;\nmandatory-level heights are obtained by using the hydrostatic equation\nbetween sigma surfaces.","-6-\nUntil May 1975 the potential temperature and wind were assumed\nto vary linearly with the Exner function TT = (p/1000)k; the mean of\nthe values of TT at the adjacent levels was assigned as the layer pressure.\nThese assumptions introduced a substantial bias in the stratospheric\nheights and temperatures, a bias caused by the mechanics of the\ncoordinate transformation and unrelated to the dynamics of the fore-\ncasting procedures.\nA more accurate system was developed in which the temperature\n(not the potential temperature) is assumed to vary linearly with the\nnatural logarithm of pressure; the layer quantities are assigned that\nvalue of pressure which relates to the potential temperature (0) of the\nlayer (a given quantity) and to the \"thickness temperature\" for the sigma\nlayer. This \"thickness temperature\" (T) is determined by solving the\nhydrostatic equation dgz/dlnp = - RT for the sigma layer. Thus T = T/O\nyields the P value.\nThis change, introduced into the output sections of the 8L GLOBAL\n(2.5°) on May 28, brought about a reduction in the stratospheric height\nand temperature biases. The tropospheric quantities were unaltered.\n(Stackpole and Collins)\n2.1.4 Global Model Noise Suppression\nA series of experiments to test various techniques of suppressing\nnoise in the global model were concluded. Two techniques provided\nimproved noise control over the operational one which consists of a\nthree-point time smoothing (used originally by Robert) One beneficial\ntechnique consisted of damping the vertical mean divergent wind component\nin layers above the material surface in the vicinity of the tropopause.\nThe other method consisted of damping the mass-weighted vertical mean\ndivergence in this same region of the model atmosphere. When damping\nthe wind divergence throughout the entire atmosphere, precipitation was\nseriously depleted and anomalous flow was created over mountains.\nAn added advantage of these new techniques is that they permit\na longer time step than did the time filter control. A 14% computer\nsavings was realized. A report documenting these results is under\n(Dey)\npreparation.\n2.1.5 Global Model Output Code\nAn output code has been written for the 8L GLOBAL. The program\nuses sigma layer data as input. Contour maps of a number of fields may\nbe produced on any combination of three map projections: Northern and\nSouthern Hemisphere polar stereographics; mercator tropical strips.\nThe code may be used with any available latitude-longitude grid. (Dey)","-7-\n2.1.6\nTropopause Experiments\nThe current method of calculating tropopause pressures for the\n8-layer Flattery analysis had several deficiencies: errors in tropo-\npause pressures; errors in heights and temperatures at mandatory\npressure levels.\nTests replaced the upward-searching technique, used to locate\nthe tropopause region, with a downward one. Results show a significant\nreduction in the RMS error of tropopause pressures. Verification of\nheights and temperatures at mandatory pressure levels, carried out after\na pressure-sigma-pressure transformation, showed similar results.\nExperiments are underway to determine how many layers are\nrequired to adequately define the vertical structure of the model\natmosphere with and without the aid of a tropopause level. (Bostelman)\n2.1.7\nSpectral Modeling\nExperiments with the complete spectral treatments of all spatial\ncoordinates have not produced stable runs for periods of 48 hours with a\ntwenty zonal modes resolution. However, an eight-wave truncation has\nbeen integrated satisfactorily.\nIt seemed reasonable to experiment with a vertical finite\ndifferencing scheme incorporating the conservation properties suggested\nindependently by J. A. Brown and N. A. Phillips. To this end a second\nmodel version has been coded with satisfactory results up to 48 hours.\nAn interface program between the Hough-to-spherical-harmonic\nconversion and the spectral model is now being written to interpolate\nthe mandatory level analysis into the sigma layers. In the process, it\nwas found that with a choice of zero pressure at the model's top the\nhydrostatic diagnosis of temperatures produces highly unrealistic results.\nWhile a nonzero top pressure alleviates this problem, the resulting\nsystem of equations is not particularly suitable for a semi-implicit time\nintegration. Consequently, the finite difference version of the\nvertically integrated hydrostatic equation was sacrificed--thus retaining\nthe semi-implicit advantage.\n(Sela)","-8-\n3L GLOBAL Forecast Model\n2.1.8\nSeveral major revisions have been made more recently in the 3L\nGLOBAL model and its allied programs. The initial data extraction program\nwas modified to provide a more accurate return to the initial data (wind\nand height on pressure surfaces) from the model data (wind and temperature\non sigma surfaces). The model itself was converted to three layers of\nequal pressure difference in the vertical and the Brown-Phillips energy-\nconserving scheme for calculating pressure force was incorporated.\nThe error in the global analysis program was eliminated in\nApril 1975 and a new data set for late May 1975 was collected. A fore-\ncast to 4 days has been calculated from 0000 GMT May 22, 1975, initial\nanalyses. Verifications were made over various regions of the globe with\nsatisfactory results in all regions except in the tropical latitudes,\nwhere little or no forecast skill was found. The output and verification\nprograms are being modified in preparation for calculation of experimental\n3L GLOBAL model forecasts to 132 hours for evaluation for accuracy and\n(Vanderman)\nvalue to NMC's extended forecast efforts.\n2.1.9 Momentum Flux from Mountain Gravity Waves\nWork has continued, in Sweden, toward the development of a\nnumerical model of gravity waves suitable for use within the framework of\na NWP model. The equations of the wave model have been developed and\nsample calculations performed illustrating the potential utility of the\n(Collins)\nmodel.\n2.1.10 Stochastic Dynamic and Monte Carlo Forecasting\nStochastic dynamic prediction is a technique whereby statistical\nhydrodynamical models are used to forecast directly mean and variance\ninformation of an evolving ensemble of atmospheric states. The mean pro-\nvides the best estimate of the true state, in a least mean-square-error\nsense, while the predicted covariance information provides a quantitative\nestimate of the uncertainty in the forecast. The one disadvantage of the\nmethod which hampers its practical application is the rather large amount\nof computer time required. In order to alleviate this problem it may be\npossible to utilize a Monte Carlo approximation to the stochastic dynamic\nequations. Instead of forecasting the mean and covariance information\ndirectly, as is the case with the stochastic dynamic method, a relatively","-9-\nsmall number of solutions is run and the lowest order moments are\nestimated indirectly from this finite sample. Work is proceeding to\ntest this hypothesis by making use of a simple barotropic model.\nAlthough the sample statistics are unbiased, they are nevertheless\nsubject to sampling error in view of the limited sample size. Hence,\none important question to be answered relates to the proper sample size\nwhich would provide stable statistical estimates.\n(Pitcher)\n2.2\nRegional Modeling Branch\n2.2.1\nGeneral\nWork on the semi-implicit and planetary boundary layer models,\ndescribed in the 1974 Annual Report, was suspended due to higher\npriority tasks. Mr. Svante Bodin completed his visit in July and\nreturned to the Swedish Meteorological and Hydrological Service.\nTasks of limited scope were completed: a test of the sensitivity\nof LFM forecasts to a data base which excluded observations above 200 mb\ncompilation of comparative humidity analyses with and without satellite\nVTPR relative humidity estimates; production of KCRT depictions of the\nLFM predicted stability index for use by the Severe Storms Forecast\nCenter.\n(Gerrity)\n2.2.2\nLimited-area Fine-mesh Model (LFM)\nDuring the first quarter of the year, the operational LFM fore-\ncast was extended from 24 to 36 hours. A detailed study of those cases\nin which the forecast failed to reach 36 hours resulted in changes in\nthe model procedures that enabled the LFM to continue past the point of\nfailure. In particular, model instability associated with the LFM\nboundaries was avoided by imposing a maximum allowable value on the\nmagnitude of the wind components at each time step. In December, the\nroutine LFM forecast was extended to 48 hours.\nEarlier in the year, the noise level, when reviewed in the\nsigma domain of a routine LFM forecast, was quite high. Filters remove\nmuch of the noise after the sigma-level data are interpolated to\nconstant-pressure levels in post processing; a noisy forecast, however,\nwill contain spurious small-scale features in the sea level pressure or\n500-mb absolute vorticity fields, for example, in the post-processed\noutput.","-10-\nOriginally, the LFM forecast space-smoothed temperature and\npressure at each time step but did not include the two outermost rows\nof the grid in the procedure. Consequently, noise from the boundary\npropagated into the interior. Recent smoothing modifications have\nincluded all interior grid points and, in addition to temperature and\npressure, the horizontal wind components: changing the numerical value\nof a smoothing coefficient controls the noise level rather precisely;\nspace filtering wind, temperature, and pressure largely obliterate the\nshortest wavelengths but leave the longest wavelengths relatively\nuntouched. The new procedure reduces the noise-level significantly,\nand both the reliability and credibility of the model have improved.\nIn the belief that the model was overforecasting the areal\nextent of convective precipitation, the treatment of moist convective\nprocesses was examined. Although several variations in the computational\nmethod were tested, no definite conclusions were reached. The testing\nshowed, however, that these calculations accounted for 20 percent of the\ntotal central-processing time of a forecast. Several cases were rerun\nand the convective processes computed only once an hour, rather than\neach time step (6 minutes). Little difference was noted in the precipi-\ntation forecasts, but the saving in computer time was significant. This\nchange was incorporated in the operational LFM in November 1975.* (Newe11)\nVery-fine Mesh Model (VFM)\n2.2.3\nThe VFM, a version of the LFM, employs a mesh-length one-half\nthat of the LFM. Although this model covers one-quarter the area of\nthe LFM, the VFM does encompass the 48 conterminous states and southern\nCanada. The grid length varies from 95 km at 60 degrees north to about\n75 km at 30 degrees north.\nOne 24-hour forecast has been made using boundary conditions\ntaken from a previously made LFM forecast. Possibly due to a mismatch\nbetween the pressure forecasts of the two models, difficulty in predicting\nthe mass field was noted. A solution is being sought.\nIn connection with finer-mesh models, preliminary work has begun\non a modified version of the LFM which will employ smaller mesh lengths\nthan that of the operational version but cover the same geographical\narea. Both this model and the VFM should provide insight into the\n(Newe11)\nbenefits of models of higher resolution.\n*Verifications of LFM Operational Precipitation Forecasts may be found\nin Part VI, page 34.","-11- -\n2.2.4\nLFM Boundary Condition Specification\nThe limited-area prediction model's atmosphere is coupled to\nthe 6L PE atmosphere by time-dependent lateral boundary conditions.\nThese conditions are extracted from the 6L PE model forecast made at\nthe previous synoptic cycle. The models have similar vertical\nstructures permitting the boundary conditions to be specified,\ndirectly, in terms of the sigma coordinate variables.\nThe planned implementation of the 8L HEM necessitates develop-\nment of a different procedure because the sigma coordinate systems are\nincongruent. A new technique, to be tested early in 1976, converts\nthe 8L HEM sigma-coordinate forecast data into isobaric data; this\ndata is reconverted into the sigma system of the LFM.\n(Polger)\n2.2.5\nLFM Comparative Verification Statistics\nComparative verification statistics have been made between\nthe LFM and the 6L PE. These data indicate that the LFM has a clear\nsuperiority at 500 mb, particularly at 36 hours. It is not clear,\nhowever, from the S1 scores that this is the case for the surface\nforecasts. There is evidence, however, that the LFM precipitation\nforecasts serve as better guidance for use by the NMC forecasters\nthan the precipitation forecasts of the 6L PE.\nDuring the latter part of December 1975, the LFM forecasts\nwere extended to 48 hours on an experimental basis. Statistical\nevaluation of these forecasts will be compiled on a continuing basis\nthrough 1976.\nThe data presented in this section were provided by R. Van\nHaaren of Data Assimilation Branch.\n(Gerrity)\n2.2.6\nHemispheric Nested Grid Model\nThis model has been brought to the state where introduction of\nreal data can soon be started. Its format is flexible, not only in\nthe number and location of the sigma surface but also in the number of\ninteracting grids (1, 2, or 3) and their size, location, and orientation.\nComputation time estimates are of the order of an hour CPU time for a\n24-hr forecast on the IBM 360/195 for a 6-1evel version in which the\nfinest grid has an area about 2/3 the size of the present LFM grid but\nwith twice the resolution. Two model features are of possible\napplication to other NMC models. One is the development of a hemi-\nspheric orography grid which has been smoothed by a 72-wavenumber","-12-\nspherical harmonic filter. The other is the use of standard\natmospheric height--a function of pressure--in place of log pressure\nor pressure to the 2/7 power as the vertical coordinate in interpolation\nbetween sigma and pressure coordinates. This has the advantage of\nexact reversibility in the sigma-pressure conversion for any temperature\ncurve whose lapse rate is that of the standard atmosphere.\nCurrent plans are first to run the model without physics on\nseveral real cases to verify its computational behavior. Physics will\nbe added after this final stage (most likely a strict copy of the\nphysics in present operational models). (Phillips, Campana, and Mathur)\nNested Grid Hurricane Model\n2.2.7\nResults from four numerical experiments using a four-level\nnested grid PE model designed to simulate the development of hurricane\nIsbell 1964 were compared. These experiments differed from each other\neither in the formulation of total latent heat release or surface\nfriction. A cylindrical grid system (best suited to study dynamical\nstructure of hurricanes) cannot be designed to have everywhere the same\nhorizontal grid resolution as the Cartesian grid used in the model.\nIt was found necessary, therefore, to calculate such terms as horizontal\nadvection and pressure gradient at the Cartesian grid points first, and\nthen interpolate to make the coordinate transformation at the cylindrical\ngrid points. These results show that the asymmetric structure of\nsurface winds (vo) is better simulated when the drag coefficient CD\nis\nassumed to vary as CD = (0.7 + 0.7 Vol: ) 10-3 than when it is assumed\na constant (value 0.0025). Surface friction tends to decelerate the\ninflow in the hurricane core region. The vertical motion at the top of\nthe boundary layer (WB) may be visualized as associated with (1) the\nfrictional convergence in the boundary layer, and (2) the strong zones\nof convergence induced in the boundary layer by the development of\nstrong outflow regions in the upper troposphere. The latter contribution\nto WB is very large in the hurricane stage. The maximum winds in the\nboundary layer are located in the region of inflow, in contrast the\nmaximum winds in the middle troposphere are supergradient and occur in\nthe region of outflow. Some of these results were presented at the AMS\nNinth Technical Conference on Hurricanes and Tropical Meteorology in\n(Mathur)\nMay 1975 at Miami.\nAir-Sea Interaction and Convective Parameterization\n2.2.8\nA new LFM test code modifies the LFM forecast code to replace\nthe convective adjustment scheme with Kuo-type convective parameteri-\nzations. The code also incorporates four moisture levels (instead of\nthe forecast code's three) plus the sea-to-air latent heat supply.\nResults suggest that Kuo-type parameterization leads to somewhat larger\nheating in the upper troposphere. Because of the vertical staggering of\nvariables in the LFM, it is felt that warming in the upper troposphere\nmay induce too much divergence in the model stratosphere with a resultant\nloss of stratospheric mass. Stratospheric \"exhaustion\" occurs more\n(Mathur)\noften with Kuo-type parameterization.","-13-\n2.2.9\nExperimental Forecasts--Hurricane Project\nFive test cases with Hurricane Carmen early in the year showed\nsurprising accuracy, yielding 36-hr forecasts of comparable accuracy\nto operational 24-hr predictions. Tests under semi-operational\nconditions during the summer and fall essentially confirmed the early\nperformance levels and most forecasts were provided to the National\nHurricane Center within the 12 hours of initial data time. The\nfollowing table provides a breakdown of the number of prediction runs\nmade for each storm:\nStorm\nNumber\nof Runs\nAMY\n4\nBLANCHE\n3\nCAROLINE\n3\nDORIS\n0\nELOISE\n12\nFAYE\n3\nGLADYS\n6\nTOTAL\n31\nOn the average, position accuracies under semi-operational -\nconditions were not as outstanding as those produced for Carmen. This\nwas to be expected as considerable testing was performed in the vicinity\nof high-speed jets. The paths of the storms continued to be predicted\nexcellently; speed of movement, however, was poorly handled at times.\nThe major problem that surfaced during the year of testing\ninvolved the model's oversized initial vortex. This idealized storm is\nderived from a time integration of a two-dimensional prototype of the\ncomplete model. Lack of sufficient data in hurricanes has necessitated\nsuch an approach which guarantees a vortex in tune with the model\nequations but generally not a good reproduction of the atmospheric\nstorm. The relatively coarse grid resolution forces the numerical\nstorm to reach a horizontal extent larger than that generally observed","- -14-\nand this in turn results in an erroneous northward movement. The\nproblem becomes noticeable as the grid is positioned at latitudes\nbelow 20°N. Some temporary adjustments can be made to alleviate the\nproblem during the coming year, but a complete solution can be achieved\nonly by reduction of truncation error in the central portions of the\ngrid.\nExperimentation is continuing in order to better understand\nthe hydrodynamic behavior of model storms as they move in simple flows,\nas well as those in the atmosphere. The vorticity budgets around\nmoving storms are under investigation with primary interest focused on\nhow the various terms in the vorticity equation are influenced by\nchanging flow fields. The ageostrophic character of the forecasts have\nbeen evaluated in terms of storm movement as well. Simple, less\nexpensive prototypes have proved to be valuable tools in these studies.\nIn preparation for the coming season, further optimization of\nmachine codes has been carried out with reductions of approximately 10%\nin running time. Further time reductions involving data input into the\nmodel will be incorporated within the next few months.\nExperimental forecasts have also been performed in non-\nhurricane situations with initial data interpolated from the NMC global\noperational analyses. The model has demonstrated a noticeable improve-\nment over operational precipitation forecasts, showing in most cases\nbetter placement of large-scale features and an apparent ability to\npredict fine-scale features not resolvable on conventional grids.\n(Hovermale, Scolnik, and Marks)\nIsentropic Analyses--Hurricane Project\n2.2.10\nThe isentropic analysis technique has been modified to obtain\nbetter vertical resolution (30 levels are treated rather than 10). This\nhas placed more stringent requirements on data integrity and has\nincreased the need for vertical consistency between levels. Considerable\neffort has been required to write programs to thoroughly check reports\nand to employ a variational technique that provides more vertical con-\nsistency in the final analyses.\nA humidity analysis has been added to the package and evaluated\nin broad terms with fields obtained from the LFM. The codes appear to\nbe functioning properly and further testing will require more objective\nstandards.","-15-\nThe method includes an enhancement of boundary layer humidity\nfields through treatment of surface observations coupled with vertical\ngradients from radiosonde reports. Where radiosonde data is lacking,\nhumidity estimates are inferred from local surface weather and cloud\nobservations. These estimates, judiciously used, supplement the data\nemployed in the overall humidity analysis.\nAn interpolation scheme which employs elliptic influence\nfunctions has been modified to make it applicable to moisture as well\nas wind analysis. The new version of the code, applicable to any size\narray, utilizes a \"B-spline\" under tension. It has been tested in\nanalytic and real data cases with satisfactory results.\n(Marks, Chu, and Jones)\n2.3\nData Assimilation Branch\n2.3.1\nGeneral\nDuring this past year, the Branch was heavily involved in Data\nSystems Test (DST) activities for NASA. Conduct of impact tests and,\nto a lesser extent, efforts in data assimilation were limited until\nabout November by extreme difficulty in obtaining turnaround for large\ncomputer jobs on the dual IBM 360/195 system. Archiving and impact\ntests were run with almost the entire DAB staff. A new effort began\nthis year in forecast verification and model test and evaluation.\n2.3.2\nData Archiving\nLevel II data sets were generated in real-time for the NASA\nData Systems Test from 18 August through 18 October 1975. Types of data\nincluded and periods when these data were available are summarized in\nthe following table:","-16-\nPeriod C\nPeriod A\nPeriod B\n22 Sept- - 18 Oct\n18 Aug-4 Sept\n5 Sept- - 21 Sept\nType of Observation\nX\nX\nX\nSurface observations\nX\nX\nX\nRadiosonde reports\nX\nNESS cloud-tracked winds\nX\nX\nWisconsin cloud-tracked\nX\nX\nwinds\nX\nX\nOperational VTPR soundings\nX\nX\nNimbus 6 soundings\nX\nX\nX\nAircraft reports\nX\nX\nSpecial aircraft reports\nX\nConstant-density balloon\nX\nX\n(TWERLE) data\nX\nAverage daily collections during Period A include about 6000\nsatellite temperature soundings and 5000 cloud-tracked winds. Data tapes\nfor this period were mailed to Goddard Institute for Space Studies (GISS), ,\nthe National Center for Atmospheric Research (NCAR) , the Geophysical\nFluid Dynamics Laboratory (GFDL), Professor Arakawa at UCLA, and the\n(Desmarais, Chiusano, and O'Neil)\nBureau of Meteorology, Australia.\nGlobal analyses were produced from Period A Level II data sets\nusing the 8L GLOBAL (2.5°) and the Flattery spectral analysis scheme.\nAnalyses were produced twice daily with a 12-hr analysis-forecast cycle.\nTapes containing analyzed values of height, temperature, moisture and wind\n(Level III data sets) were mailed to GISS, GFDL, NCAR and UCLA for use\n(Desmarais, O'Neil, and Bonner)\nin GARP-related experiments.","-17-\n2.3.3\nData Impact Tests\n(a)\nA very limited test was conducted in which VTPR soundings\nderived at GISS were used in the NMC global analysis-forecast system.\nGISS soundings were produced at 2.50 latitude-longitude intersections\n(grid points of this global model) from radiance measurements supplied\nby NESS and the NMC global 12-hr forecast temperature profiles.\nAnalyses were produced from 12 GMT 4 February to 00 GMT 9 February\n975--using the normal NMC data base, except that GISS VTPR soundings\nreplaced the operational retrievals. GISS soundings were given the\nsame weight as radiosonde data; resulting analyses were excessively\nnoisy and showed small-scale features that were not realistic. Height\nfields at 1000 mb and temperatures aloft were frequently inconsistent--\nat least partly because the GISS soundings were provided to us with\n1000-mb heights that were actually 12-hr forecasts from our analysis-\nforecast cycle. Further tests of \"high density\" retrievals will be\nconducted using Nimbus 6 data.\n(Desmarais and Bonner)\n(b)\nA fairly extensive test of the impact of satellite temperature\nsoundings was conducted using DST data sets collected in August and\nSeptember (see previous section). DST Level III analyses were compared\nwith analyses produced by a parallel system using the same data base but\nwith satellite soundings (VTPR and Nimbus 6) removed. A series of 84-hr\n6L PE forecasts were made from DST and parallel mode analyses on 10\nforecast days. Results showed, once again, little or no improvement in\nforecast skill from use of satellite temperatures.\n(O'Neil, Bonner, and Desmarais)\nA third analysis-forecast cycle, run for only 5 days, used the\nfull DST data set with radiosonde observations removed. Temperatures\naloft were provided only by satellite soundings. Analyses generated in\nthis mode agreed in all major features with the operational NMC\nanalyses on corresponding days; however, as in an earlier experiment,\nthe analyses contained systematic underestimates of the intensities of\nmajor systems and the strength of the wind. At least some of this under-\nestimation is to be expected from use of asynoptic data in a 12-hr fore-\ncast cycle.\n(Bonner, Desmarais, and O'Neil)\nA condensed version of the NASA Study Report describing the\nVTPR impact test (see NWP Activities Report 1973) was submitted for\npublication as a NOAA Technical Memorandum.\n(Bonner, O'Neil, Desmarais, Van Haaren, and Lemar)","-18-\nAn article describing the ITPR impact test (see NWP Activities\nReport 1974) has been submitted for publication as a NOAA Technical\n(Bonner, Van Haaren, and Hayden (NESS))\nMemorandum.\nTemperature and moisture data from Nimbus 6 are being\n(c)\nevaluated through construction of cross-sections and constant-pressure\nanalyses in a series of case studies from the August-September DST\n(Tracton)\nperiod.\nFairly extensive comparisons are being made between rawinsonde\n(d)\nreports and co-located mid- and high-level satellite winds generated\nby the University of Wisconsin during the August-September DST. Results\nof an earlier, similar study (see NWP Activities Report 1974) were\n(Lemar)\ndocumented in NMC Office Note 114.\nModel Test and Evaluation\n2.3.4\nWork was begun on the development of a standard verification\npackage for routine, operational evaluation of the LFM, 6L PE, and\n8-1ayer model forecasts. Specialized verification and diagnostic\nroutines are being developed for evaluation of the ability of NMC\nmodels to forecast cyclogenesis. As DST activities subside, the Branch\nwill assume major responsibility for model evaluation within NMC.\n(Van Haaren, Tracton, and O'Neil)\n2.3.5\nData Assimilation\nTwo parallel experimental global assimilation systems based\non different horizontal resolutions of the NMC 8L GLOBAL are being\ntested. The goal of this effort is to develop an assimilation system\ncapable of being used operationally during the First GARP Global\nExperiment (FGGE).\nSystem I. This system involves simply a reduction of the NMC\n(a)\nFINAL 12-hr global analysis-forecast cycle to a 6-hr cycle in which\ndata are treated as synoptic within a window of + 3 hrs. A parallel\ntest of 6- and 12-hr cycles with the spectral analysis scheme and\n8L GLOBAL (2.5°) model conducted during 1975 revealed numerous problems\nwith the 6-hr cycle. Most of these problems were related to the\naccumulation of noise in the system. Apparently, 6 hours is not\nsufficient time to dissipate the \"shock\" introduced by the analysis\nprocedure and the transformation from p to o coordinates. Efforts are\n(Desmarais, Rasch, and McPherson)\nunderway to overcome this problem.","-19-\nA series of sensitivity experiments are being conducted to\nproduce a better understanding of the characteristics of the spectral\nanalysis scheme. Particular attention is being given to the relation-\nship between winds and heights and to the manner in which single\nlevel reports influence analyses at other levels. A report is being\nprepared describing the results of these experiments. Several memos\nwere written describing various features of the spectral analysis\nscheme. These memos are the first step in preparation of a report\ndocumenting the overall features of the analysis code.\n(Rasch)\nOn August 7, 1975, blending of height and wind analysis co-\nefficients was eliminated from the FINAL analysis. This was done to\neliminate unrealistic warming near the poles and cooling near the\nEquator in the stratosphere, which results from not including the\ncyclostrophic component in the wind law which is enforced by the\nblending. However, large ageostrophy can exist now in data-void\nareas of the analysis. Experimentation is now underway to re-\nintroduce the blending with the cyclostrophic effect included.\n(Parrish (Automation Division))\n(b)\nSystem II. The second system differed from the first\nprimarily in that updating of the prediction model was done by means\nof a local interpolation method, affecting only grid points in the\nvicinity of observations. In addition to reducing the shock of data\ninsertion and thus the cumulative noise level, this also permits\ngreater flexibility in treating observations of nonhomogeneous quality\nand characteristics. System II will ultimately use a multivariate\nthree-dimensional optimum interpolation procedure.\nIn the prototype constructed during 1975, however, a two-\ndimensional Cressman-type analysis of temperature and surface pressure\nwas used to update the 5° version of the NMC 8L Global model (8L GLOBAL\n(5°)). Preliminary comparative tests of both systems have shown that\nlocal updating results in a smaller shock following data insertion\nthan was evident in System I. Testing of both systems (using 5 o\nresolution) will continue.\n(Kistler, Gordon, and McPherson)\nAn optimum interpolation analysis program for surface pressure\nwas developed and tested. The analysis results compare favorably with\nanalyses produced by the Cressman and Flattery methods.\n(Bergman and Kistler)","-20-\nA single grid point version of the multivariate optimum\ninterpolation code was written and is now ready for testing with\nartificial data. Some modifications to the analysis system as\noriginally described in Office Note 116 were made. These include a\nrevised treatment of the polar-latitude forms of the analysis\nequations, the use of a latitudinally varying function to decouple\nthe cross-correlations between temperatures and winds, and a change\nin the vertical correlation functions used in the wind component\nanalyses. A univariate analysis of dew-point depression and an\nerror-checking routine which utilizes optimum interpolation will be\n(Bergman and Gordon)\nincluded in the final analysis package.\nForecast error covariances and correlations required by the\noptimum interpolation scheme were computed for 12 mandatory pressure\nlevels using worldwide rawinsonde data from 10 randomly selected\nsynoptic times. Results of these computations suggest the change in\nthe vertical correlation functions mentioned above. Horizontal and\ncross-correlations will be computed using the same rawinsonde data\n(Gordon and Bergman)\nbase.\n2.3.6\nRelated Projects\nA paper \"Analysis Error as a Function of Observation Density\n(a)\nfor Satellite Temperature Soundings with Spatially Correlated Errors,\"\noriginally issued as Office Note 119, has been revised and submitted\nto the Journal of Applied Meteorology for publication. The revised\npaper includes a tentative error correlation function for neighboring\nsatellite temperature observations. This function was obtained for a\nlimited data sample by comparing Nimbus satellite temperatures with\ncross-section analyses based on rawinsonde data.\n(Bergman, Bonner, Lemar, and Van Haaren)\nAn examination of the information content of VTPR retrievals\n(b)\nwas carried out during 1975. In this study, co-located VTPR and radio-\nsonde temperature and thickness profiles were fit by squares to\nappropriate sets of empirical orthogonal functions. RMS differences\nwere computed as a function of the number of empirical functions used\nto represent the profiles. It was found that reducing the number of\nfunctions from 12 (perfect fit) to 5 resulted in only a slight decrease\nin the RMS difference. Thus, it appears that most of the RMS difference\nof 2 to 3 degrees between co-located VTPR and radiosonde reports is\nassociated not with differences in the vertical resolution of the two\nsounding systems but with variations in \"true\" temperature between\n\"co-located\" reports and with large-scale vertical bias in the VTPR\n(McPherson and Gordon)\nretrievals.","-21-\n(c)\nDocumentation of a series of experiments performed in 1974\ninvestigating the performance of various insertion techniques within\nthe content of a primitive-equation barotropic model was completed\nduring 1975. A geostrophic correction to the motion field based on\nobservations of the mass field was reported in a paper entitled\n\"On the Use of a Local Wind Correction Technique in Four-Dimensional\nAssimilation\" (Kistler and McPherson, MWR, 103, No. 5) Character-\nistics of several damping time integration methods and their\napplication to data assimilation were examined in NOAA Technical\nMemorandum NWS NMC-56 (McPherson and Kistler). Finally, NMC Office\nNote 106 describes experiments in which data sets were inserted\nrepeatedly in an effort to improve the assimilation of the data.\n(McPherson and Kistler)\n(d)\nA review paper entitled \"Progress, Problems, and Prospects\nin Meteorological Data Assimilation, was published in the November\nissue of the Bulletin of the American Meteorological Society.\n(McPherson)\n2.4\nUpper Air Branch\n2.4.1\nAnalyses of Stratospheric Data\n(a)\nRevision of stratospheric analysis program (70-10 mb) . After\nstudies were made of satellite data in comparison with neighboring\nradiosonde data, the analysis code was revised to permit the analysis\nof VTPR height and temperature data at selected levels. Beginning on\nJuly 22, 1975, satellite data were included operationally at all four\nanalysis levels, though heavy smoothing and tight scan-limits somewhat\nlimited their overall effect. Later on, the VTPR data were removed\nfrom the original (daily) version of the analysis program to maintain\nconsistency with the NMC FINAL analysis during the DST period. (A\nfurther consideration was the increasing incompatibility between VTPR\nand radiosonde data as winter approached.) New program changes (12/18)\nincluded adjusting height and wind weighting, tuning toss-out limits\n(especially at low latitudes), and replacing heavy smoothing with a\n9-point filter. Results revealed better RMS fit of analysis to data,\nand exposed problems with regressions and data-handling.\n(Laver and Finger)","-22-\nRocketsonde-satellite comparisons. United States rocket\n(b)\nstations have been scheduling meteorological rocketsonde- launches to\ncoincide with NOAA and Nimbus satellite overpasses. The radiances\ncomputed from the rocketsonde-radiosonde temperature profiles are\ncompared with the satellite radiance measurements in the tropospheric\nand stratospheric channels. Results from comparisons for NOAA 3 and\nNOAA 4 VTPR data have indicated a correlation between biases in the\nradiance measurements and biases in temperature retrievals from 1000\nto 10 mb. The National Environmental Satellite Service has switched\nto a system of regression retrievals (based on coincident radiosonde\nobservations) to minimize the effects of biased differences from\nradiosonde measurements.\nRocketsonde comparisons with Nimbus 6 sounders (HIRS, LRIR,\nand PMR) launched in June 1975 have begun. These comparisons will be\nespecially useful for determining the compatibility of the various\nstratospheric and mesospheric sounders. (Gelman, Miller, and Finger)\nInternational rocketsonde comparisons. Comparisons were made\n(c)\nof the temperature and wind information obtained from the rocketsonde\nsystems of France, Japan, the United Kingdom, the Union of Soviet\nSocialist Republics, and the United States. Evaluation of the results\nallowed adjustments to be derived, leading to more meaningful use of\nthe data for depicting stratospheric-mesospheric circulation. A\npaper describing the results was published in the Journal of the\n(Finger and Gelman)\nAtmospheric Sciences, September 1975.\nRocketsonde data-exchange and analysis. Rocketsonde data are\n(d)\nused on a continuing basis in weekly meridional cross-sections from\n20 to 70 km. These analyses are being exchanged with the Soviet Union\nunder a bilateral agreement for cooperation in space research.\n(Finger, Gelman, McInturff, and Nagatani)\nEvaluation of satellite data. Mapped fields of VTPR strato-\n(e)\nspheric channel radiation data were monitored, and feedback provided\nto NESS on instrumental irregularities. Scan-bias problems in retrieval\ntemperatures at 10 mb were reported, resulting in regression modifi-\ncations by NESS. Preliminary comparisons were made of HIRS-derived\nstratospheric temperatures with independent data. Evaluation of\nBritish PMR radiation fields has been initiated.\n(Quiroz, Gelman, and Nagatani)\nFalling-sphere soundings. Report on reliability of temperatures\n(f)\nfrom sphere drag soundings, needed for checkout of new satellite\ninstruments (PMR, LRIR) has been completed; to be published in Journal\n(Quiroz and Gelman)\nof Geophysical Research.","-23-\n(g)\nSST forecasts. Reevaluation of NMC's 100-mb and 70-mb fore-\ncasts (SST flight levels) is tentatively planned for the winter of\n1975-76, using latest version of the NMC 8-1ayer Global Model.\n(Laver, Nagatani, and McInturff)\n(h)\nComparison of Flattery spectral and Cressman-type analysis\nin stratosphere. Preliminary comparisons of these two systems at 70\nand 50 mb were begun. Estimates are being made of accuracy, effects\nof VTPR data, and feasibility of using the spectral analysis system\nat stratospheric levels.\n(Myers and Laver)\n2.4.2\nResearch on Atmospheric Circulation\n(a)\nStratospheric warmings and polar vortex breakdown. Papers\non (1) evolution of warmings, based on analysis of satellite radiation\ndata, and (2) comparative features of warmings, simulated and observed,\nwere published in the Journal of Atmospheric Sciences. Wintertime\ndisturbances of Northern and Southern Hemisphere stratospheres were\ncontinuously monitored. Analysis of radiation fields, construction of\ncirculation maps, investigation of relevant tropospheric parameters,\netc., , were conducted.\n(Quiroz, Miller, Nagatani, and others)\n(b)\nStratospheric warmings and tropospheric blocking. Preliminary\nefforts are being made to correlate stratospheric warming trends with\ntropospheric blocking. Height anomaly patterns are being studied\nduring periods of stratospheric warming activity.\n(Myers, Laver, and Finger)\n(c)\nGlobal energy program. The program for computing several\nterms of the global energy balance has attained operational status.\nIn addition, the program is being run off the 6L PE forecasts at 00,\n12, 24, 36, 60, and 84 hours, and it is planned to evaluate the 8-1ayer\nglobal forecasts in a similar manner. (Miller, Collins of Global Mod. Br.)\n(d)\nSTRATWARM monograph. An assessment is being undertaken of\ncurrent knowledge of the stratospheric-warming phenomenon. It takes\nthe form of a complete description from the synoptic, dynamic, and\nenergetic points of view. Particular attention is given to interactions\nbetween the stratosphere and troposphere, and between the stratosphere\nand mesosphere. First-drafts of all chapters have been completed.\n(Upper Air Branch Staff)","-24-\nPART III TECHNIQUES DEVELOPMENT AND APPLICATION OF NEW PRODUCTS\n(Systems Development Office, Techniques Development Laboratory)\nGeneral Development Techniques\n3.1\nThe forecast research program of the Techniques Development\nLaboratory (TDL) is divided into public weather prediction, severe\nlocal storm prediction, marine environmental prediction, terminal\nweather prediction, and AFOS* forecast applications. In each of these\nmain areas, the principal goals are to develop better objective methods\nfor making weather forecasts and to provide for their implementation at\nthe appropriate place in NWS.\nIn order to accomplish these goals, TDL collects extensive\nsamples of forecasts from operational numerical models and statistically\nrelates weather variables such as temperature, winds, precipitation,\nand occurrence of thunderstorms to these forecasts. This technique is\ncalled Model Output Statistics (MOS). New numerical models are also\ndeveloped when the need arises. For instance, a comprehensive model\nof the boundary layer is being built which will soon fill a gap in\nthe spectrum of NWS models. Initialization for numerical models\nrequires intensive study of analysis methods like successive approxi-\nmation, optimum interpolation, and variational adjustment.\nAFOS will provide a new dimension to the use of automated\ntechniques. For the first time, communication, computer, and display\nfacilities will exist at a station, which will allow rapid response to\nurgent queries, by forecasters. For instance, a forecast of precipita-\ntion can be updated very quickly based on recent radar data. Consequently,\nsome of TDL's techniques are being tailored for use on the minicomputers.\n(Klein)\nPublic Weather Prediction. Much emphasis is being placed on\n3.1.1\ndeveloping new and improved automated predictions of all weather\nelements contained in public weather forecasts. Efforts are focused\non the four key meteorological elements: precipitation, temperature,\n(Glahn)\nclouds, and wind.\n*Automation of Field Operations and Services.","-25-\n3.1.1.1 Precipitation forecasting. The automated system that produces\nnationwide forecasts of conditional probability of frozen precipitation\n(PoF) underwent extensive verification and additional study during 1975.\nNew equations based on additional data were derived for use during the\n1975-76 winter. TDL completed an experiment in which forecasts from an\nexperimental PoF system developed with the regression estimation of\nevent probability (REEP) techniques were compared with operational PoF\nforecasts made with the logit model. The results indicated that the\nlogit model should continue to be used in the operational system.\n(Bocchieri and Gilhousen)\nQuantitative precipitation forecasts (QPF) made from region-\nalized equations were supplied to NMC. Previously, only one equation\nfor the coterminous United States was used. For the winter season of\n1974-75, six regions were used. Also, categorical forecasts of pre-\ncipitation amount are now being computed from the probability forecasts\nin such a way as to maximize the threat score, a statistic used at NMC\nfor verifying forecasts of precipitation amount.\n(Bermowitz and Zurndorfer)\n3.1.1.2 Surface temperature forecasting. During 1975, TDL continued\nto monitor and improve the MOS temperature guidance. The MOS system\nforecasts maximum and minimum temperatures for 228 stations in the\nconterminous United States. These forecasts, prepared after 0000 GMT\nand 1200 GMT, use predictors from both the PE and trajectory models.\nIn July, a new series of summer regression equations based\non 5 years of data were implemented. These equations are for a shorter\nseason than used previously. A larger number of potential predictors\nwere also provided to the screening program. Recent verification has\nshown that use of the shorter season, additional dependent data, and\nthe added predictors improved the MOS forecasts. TDL intends to derive\nsets of operational equations for each of the four 3-month seasons.\n(Hammons and Dallavalle)\n3.1.1.3 Cloud forecasting. During the past year, TDL developed and\nimplemented a system to forecast cloud amount. Automated MOS forecasts\nfor 233 stations are now on teletypewriter on a request basis. Proba-\nbilities of clear, scattered, broken, and overcast, and a single \"best\"\ncategory are produced for each of seven projections (12-48 hr). The\ninput data for these forecasts are from NMC's PE model and TDL's\ntrajectory model. Separate summer (April-September) and winter\n(October-March) season forecast equations are available for each\nstation.\n(Carter)","-26-\n3.1.1.4 Surface wind forecasting. TDL derived a new set of surface\nwind forecasting equations for the summer season. Several improvements\nwere made, including increasing the dependent data sample, adding new\npredictors, increasing the number of terms per equation, and removing\nthe forcing constraint on the first three predictors. In order to\nreduce the tendency of the regression equations to underforecast\nstrong winds, a technique called inflation is now being used. This\ncauses the variability of the forecasts to be about equal to the\nvariability of the verifying observations. These changes should\nimprove the overall performance of the automated wind forecasts.\n(Carter)\nSevere local storms. One of the major research efforts in\n3.1.2\nTDL is addressed to the development of automated techniques for fore-\ncasting severe local convective weather, notably thunderstorms and\ntheir manifestations like hail, strong wind gusts, and tornadoes.\nThe forecasts cover three time ranges: 6-24 hr (medium range), 2-6 hr\n(short range), and 0-2 hr (very short range) Another task, having\na direct bearing on both the short- and medium-range tasks, aims at\ndeveloping a predictive numerical model of the planetary boundary\n(Alaka)\nlayer.\n3.1.2.1 Medium-range forecasting. In the area of medium-range fore-\ncasting, TDL has developed new multiple regression equations to predict\nthe probability of both general and severe thunderstorms 24 hours in\nadvance. The equations were obtained by relating MOS predictors to\nmanually digitized radar (MDR) data and severe storm reports. MDR\ndata are coded for blocks 40-45 miles on a side. Thunderstorm and\nsevere storm probabilities are forecast for each block in the MDR\ngrid array. The forecasts are valid for a 6-hr period centered at\n0000 GMT. The predictand for the general thunderstorm equation was\nbased solely on MDR intensity values. Different equations were developed\nfor the spring and summer seasons. The predictand for the severe storm\nequations was based on both MDR data and severe storm reports. The\nforecast probabilities for the MDR grid area are transmitted by facsimile\nto field offices of NWS, including NSSFC in Kansas City.\nAdditional efforts in the medium-range task include improve-\nments to the TDL trajectory model to provide more accurate forecasts for\nin severe storm prediction. One improvement involves the simulation\nuse of the thermal interaction between the earth's surface and the predictand\n(Reap, Kemper, and Foster)\nair parcel trajectories over land.","-27-\n3.1.2.2 Short-range forecasting. The aim of this task is to develop\nobjective techniques for making 2-6 hr forecasts of both severe local\nweather and general thunderstorm occurrences. Equations for making\nthese forecasts were developed by multiple screening regression, the\nindependent variables being selected from a set of 26 diagnostic\nquantities derived from hourly surface observations and 500-mb\ntemperature forecasts obtained from NMC's LFM model. For the 1975\nstorm season, TDL developed and implemented new regression equations\nbased on a larger data sample.\n(Charba)\n3.1.2.3 Very short-range forecasting. Effort in this range (0-2 hr)\nrelies on the capability of weather radar to identify and trace the\ndevelopment of severe local storms. Basic data consists of radar\nreflectivities conveniently digitized into nine intensity levels. The\ngoal is to develop computer programs which will identify, track, and\nextrapolate positions of these echoes. Three echo-tracking techniques\nof different complexity are being investigated. The input data to\nthese techniques consist of zero degree radar reflectivities and\npatterns of vertically integrated liquid water content (VIL).\n(Elvander)\n3.1.2.4\nBoundary-layer model. Since severe local storms are\nstrongly influenced by conditions in the lower atmosphere, TDL is\ndeveloping a three-dimensional boundary layer model (BLM) to aid in\nsevere storm prediction. This model will forecast wind, temperature,\nand humidity in the atmosphere's lowest 2 km. Initially, forecasts\nwill be for the portion of the United States east of the Rocky Mountains.\nThe grid spacing over this area is roughly 80 km. In the vertical,\neight levels with separation increasing with elevation are planned.\nParticular attention is being focused on the surface and the\natmosphere's lowest 50 m. At the surface, the model predicts temperature\nand humidity using an energy balance. Latent, sensible, and soil heat\nfluxes are constrained to equal the net radiation. Soil moisture for\nthe model will be updated daily by measured rainfall. Also, measured\nsurface temperatures will influence the model's initial soil temperature.\nWind and temperature profiles in the lowest 50 m depend on the\natmospheric stability in that layer and on surface roughness.\n(Shaffer, Yu, and Kemper)\n3.1.3\nMarine environmental prediction. In the area of marine\nenvironmental prediction, TDL has continued working to develop techniques\nfor forecasting the state of the marine environment in the oceanic,\ncoastal, and Great Lakes area.\n(Pore)","-28-\n3.1.3.1 Oceanic forecasting. Since 1968, TDL has produced automated\nwave forecasts twice daily for projections out to 48 hr for the North\nAtlantic and North Pacific Oceans. These wave forecasts are based on\nthe output from NMC's PE model. In 1975, these programs were converted\n(Pore and Richardson)\nfor use on NOAA's IBM 360/195 computer.\n3.1.3.2 Coastal forecasting. TDL implemented an automated method to\nforecast surface wind at eight offshore light stations along the east\ncoast. The MOS technique was used to produce regression equations\nfrom 3 years of light station observations and PE forecasts. Different\nforecast equations were derived for the summer and winter seasons and\nfor each of the two daily forecast times of 0000 GMT and 1200 GMT.\n(Feit)\nThe model forecasts to 42 hr in advance at 6-hr intervals.\nBoth versions of the SPLASH (Special Programs to List\nAmplitudes of Surges for Hurricanes) program--SPLASH I for landfalling\nstorms and SPLASH II for storms of general motion--were released to\nthe National Hurricane Center for operational use on the IBM 360/195\nsystem. A significant improvement was made in these programs in FY\n1975 with the derivation of a sheared coordinate system. This system\naccounts for curvature of the coastline if the curvature is not too\ndrastic. A version of SPLASH with the sheared coordinate system will\neventually replace SPLASH II for operational use.\n(Jelesnianski, Barrientos, and Chin)\n3.1.3.3 Great Lakes forecasting. TDL continued to produce operational\nwind forecasts for all five Great Lakes. The lakes are divided into 12\nforecast areas for this purpose. Wind forecasts are produced to 36 hr\nwith forecast equations based on the MOS system. Input to the forecast\nequations are numerical weather predictions from the PE model of NMC.\nDuring 1974, TDL began work on an automated wave forecast system that\nuses these objective wind forecasts as meteorological input. The\npreliminary version of the model was completed and put into operational\nuse in January 1975. Modifications to the Bretschneider wave-forecast\nmethod and Saville method of fetch length correction for limited fetch\nwidth are being used. Forecasts are made twice daily for 64 points and\n(Feit and Pore)\nare transmitted via teletypewriter.\nWork has begun on a two-phase effort to improve the objective\nforecasts of mesometeorological phenomena in the Great Lakes region.\nIn the first phase, an objective technique to analyze the 3-dimensional\natmospheric structure around and over the lakes is being developed.\nThis technique is based on a numerical variational analysis scheme that","-29-\nlinks surface and upper air observations with the atmospheric dynamics\nto arrive at a mesoscale picture of the Lakes region. In the second\nphase, statistical, short-range (6-12 hr) forecast equations using\nthe analyzed parameters as predictors will be developed. The forecast\nequations will predict such phenomena as low-level winds, temperatures,\ncloud conditions, and lake-effect snowstorms.\n(Grayson)\n3.1.4\nTerminal weather prediction. With the cooperation of the\nU.S. Air Force, TDL began operational production of objective medium-\nrange terminal forecasts. Two types of products are produced: three-\ncategory combined ceiling and visibility probability forecasts, and\nfive-category ceiling and five-category visibility probability fore-\ncasts.\nTDL developed the prediction equations for both types of\nforecasts through use of the MOS system; equations use predictors from\nthe PE and trajectory models, and surface observations 6 hr after model\nrun time. The three-category equations were developed with the single-\nstation approach; the regionalized approach was used to develop the\nfive-category equations for 14 regions covering the coterminous United\nStates.\n(Crisci, Globokar, and Hebenstreit)\n3.1.5\nAFOS forecast applications. During 1975 TDL had three active\ntasks: the automatic monitoring and updating of aviation terminal\nweather forecasts, the generation of computer-worded public forecasts,\nand the updating of probability of precipitation (PoP) forecasts.\n(Lowry)\n3.1.5.1 Aviation monitoring and updating. This task aims at\nmonitoring forecasts and updating guidance as needed. The goal is to\ndevelop a system which will automatically monitor aviation terminal\nforecasts to insure their validity and representativeness. This will\nbe accomplished with AFOS minicomputer software that will compare each\nobservation of ceiling and visibility at a terminal with its forecast\nand determine if a problem exists. If a problem does exist, the system\nwill advise the forecaster of the circumstances with a terminal alerting\nprocedure (TAP) message. The TAP message will also provide a guidance\nforecast to help in preparing a new official terminal forecast.\nIn 1975, development of the needed software began. Single-\nstation prediction equations were developed which will be used to\ngenerate the guidance forecasts contained in the TAP message. The\nequations are for six categories of ceiling and visibility for projections\nof 1, 2, 3, 4, and 6 hr.\n(Crisci)","-30-\nComputer-worded forecasts. The purpose of this task is to\n3.1.5.2\nprepare a worded forecast message, entirely by computer, which is in\nthe same form as the final product. MOS forecasts of several weather\nelements are input to the computer program. The program is running\nin a test mode once daily. Such a forecast should be a good starting\npoint for the forecaster in preparing his public weather forecast.\nIf he finds no reason to alter the forecast, he can merely push a\nbutton on the AFOS console and the dissemination will be automatic.\nLikewise, he can make a slight alteration or even change the forecast\ncompletely. The former he can do by simple text editing; the latter\nby entering a new set of forecast values and instructing the computer\n(Glahn)\nto generate a new worded forecast.\n3.1.5.3 PoP updating. TDL started this task in 1975; it is aimed\nat updating PoP guidance. The plan is to use MDR reports in a scheme\nthat leads to update PoP forecasts for the near future. MDR data\nprovide frequent information concerning rainfall on a subsynoptic\nscale. The current guidance product, based on PE model runs from\n0000 and 1200 GMT data respectively, is available by 0700 and 1900 GMT.\nMDR data are available hourly for use prior to the forecast release\ntime. Two efforts are underway; one under contract and one in-house.\nThe contract effort uses MDR data at two times, 2 hr apart,\nin order to identify and track echo groups. The movement of these\ngroups is then extrapolated several hours into the future. Information\nlike the time it takes an echo group to pass by a station and the\nmaximum intensity of that group is being recorded for regression\nanalysis. The in-house effort also used MDR data but in a different\nway. Backward trajectories are constructed from observing stations\nby using an advecting wind derived from PE forecasts. The radar\ninformation is then saved from the point of origin as determined by\nthe trajectory. These data are being used as predictors for regression\n(Gilhousen)\nanalysis.","-31-\nPART IV ANALYSIS-FORECAST SYSTEMS IN OPERATIONAL USE IN 1975\n4.1\nThe basic operational numerical weather prediction system\nremained unchanged during 1975. The four basic cycles, RADAT, LFM,\nOPERATIONAL and FINAL, were run from the 0000 and 1200 GMT data times\ndaily. These are summarized below:\nRADAT - This is a quick-look forecast initiated 1 hour and\n30 minutes after data time. It consists of the Flattery spectral\nanalysis and the filtered barotropic model forecast with long-wave\nstabilization control. The nonlinear balance equation was used in the\ninitialization. Terrain and friction effects were incorporated through\nthe use of an 850-500 mb thickness forecast made simultaneously. 500-mb\nforecasts were made to 48 hours on the 1977-point stereographic 381-km\ngrid.\nLFM - This cycle began approximately 2 hours after observation\ntime and was basically the same as was run in 1974. The major exception\nto this was the change in the forecast length from 24 hours to 36 hours.\nThe model changes which were implemented are described under the\nRegional Modeling Branch activities in Part II of this report. The\nanalysis method was a successive correction technique and the forecast\nmodel was basically the 6-layer Shuman-Hovermale system on a 53 X 45\nhorizontal grid on a polar stereographic projection true at 60°N. The\ngrid distance was 190.5 km at 60°N and the time step was 6 minutes.\nOPERATIONAL - This cycle was initiated 3 hours and 20 minutes\nafter observation time. No significant change was made in it during\n1975. It consisted of the Flattery spectral analysis technique and\nthe 6-layer Shuman-Hovermale model (6L PE) The forecast grid was\na\n65 X 65 domain covering the Northern Hemisphere with a 381-km grid\ndistance. The time step was 10 minutes. Forecasts were made to 48\nhours from the 1200 GMT data and to 84 hours from the 0000 GMT data.\nThe barotropic model was used to extend the forecasts to 96 hours at\n1200 GMT and 156 hours at 0000 GMT. As in 1974, the Reed model (1963,\nNMC Technical Memorandum NWS NMC-26) was used to produce a 1000-500 mb\nthickness prediction from 84 to 156 hours at the 0000 GMT time.\nAbout 5 hours and 30 minutes after observation time, a\nTrajectory Model designed by the Techniques Development Laboratory\n(NWS Technical Procedures Bulletin No. 97, September 1973) was run\nusing 6L PE forecast parameters. Model Output Statistics (Klein and\nGlahn, Bulletin of the American Meteorological Society, Vol. 55, No. 10,\nOctober 1974) were obtained using predictors from the 6L PE and the","-32-\nTrajectory Model. Short-range predictions were obtained for maximum\nand minimum temperatures, probabilities of precipitation, quantitative\nprecipitation amounts, cloud amounts, ceilings and surface visibilities\nand winds, and probabilities of thunderstorms and severe weather for\nseveral hundred United States and a few Canadian stations. For marine\nrequirements, forecasts were also made for surface wave and swe11\nconditions for the North Atlantic and North Pacific Oceans and for\nstorm surges along the east coast of the United States and the Great\nLakes.\nFINAL - In order to bring late arriving data into the\nnumerical system, a FINAL cycle was initiated 10 hours after observation\ntime. The data were analyzed with the Flattery analysis program, and a\n12-hour forecast was obtained with the Stackpole 8-1ayer global model on\na 21/2 latitude-longitude grid (8L GLOBAL (2.5°)). This was then used as\nthe first guess of the analyses at the next RADAT, OPERATIONAL, and\nFINAL cycles.","-33-\nPART V\nPLANS FOR FUTURE OPERATIONAL SYSTEMS\n5.1\nA major change is planned in the numerical guidance cycles\nin order to make the LFM products available to the field forecasters\nat an earlier time. During 1976 it is planned to have the teletype\ntransmission of the United States upper air observations (Part A)\ncompleted in time to initiate the LFM cycle 1 hour and 40 minutes\nafter observation times. The forecast period of this model will be\nextended from 36 to 48 hours. Forecasts using Model Output Statistics\nwill be made in which the 6L PE and the LFM parameters are used as\npredictors. These changes will require further changes to be made in\nthe RADAT cycle, the alternative to which has not yet been decided.\n5.2\nPresent plans are to replace the 6L PE in the OPERATIONAL\ncycle with a hemispheric version of Stackpole's 8-1ayer model on a\n2° latitude-longitude grid (8L HEM (2°)). The final design of the\nsystem will depend on the results of extensive testing in 1976.","7.3\n9.6\n11.6\n11.1\n11.4\n9.3\n7.3\n9.6\n11.6\n12.6\n9.9\n12.1\n15.8\n11.1\n11.4\n9.3\n19.8\n21.9\n19.0\n15.4\n18.1\n23.1\n14.3\n15.2\n18.0\n19.0\n17.3\n15.3\n17.2\n20.7\nW\nPERS\n50.6\n68.6\n70.4\n67.7\n54.6\n40.9\n55.5\n75.3\n56.4\n72.3\n100.6\n62.2\n60.4\n50.0\n38.7\n126.2\n107.0\n83.2\n105.5\n141.2\n87.2\n88.7\n75.0\n111.6\n111.3\n95.1\n79.3\n99.7\n129.9\n118.9\nH\n6.5\n7.6\n10.3\n9.6\n7.8\n6.8\n7.9\n8.9\n6.9\n7.8\n9.5\n8.6\n8.1\n6.5\n6.8\n14.4\n12.3\n10.6\n11.8\n13.9\n10.6\n10.1\n8.3\n15.5\n14.6\n12.8\n11.5\n12.3\n13.9\n14.9\n48 hours\nW\n36.0\n40.9\n52.7\n45.7\n37.5\n33.5\n39.9\n45.1\n75.0\n62.8\n55.5\n44.5\n38.1\n46.0\n53.0\n48.5\n42.4\n33.2\n30.5\n74.1\n62.8\n56.1\n66.5\n62.2\n58.5\n66.8\n77.4\n82.3\nPE MODEL\n85.4\n72.0\nH\n.75.\n.76\n.82\n.76\n.79\n.76\n.68\n.82\n.82\n.76\n.79\n.85\n.74\n.78\n.79\n.71\n.78\n.83\n.82\n.76\n.79\n.85\n.76\n.81\n.75\n.81\n.79\n.74\n.78\n.82\nR\n9.2\n20.5\n6.4\n7.2\nW\nBAROTROPIC\n52.7\n77.7\n36.6\n42.4\nH\n.76\n.62\n.67\n.75\nR\n9.0\n7.6\n10.9\n11.0\n8.7\n6,9\n9.1\n11.1\n9.4\n11.6\n9.0\n11.0\n14.6\n9.2\n7.3\n6.0\n19.2\n20.8\n17.6\n14.1\n16.8\n21.4\n13.9\n14.4\n17.5\n18.0\n16.2\n14.1\n15.8\n19.2\n36 hours\nW\n36.2\nPERS\n49.1\n50.3\n68.3\n44.8\n61.3\n64.0\n61.9\n65.9\n49.7\n63.7\n88.1\n56.1\n54.3\n43.9\n34.5\n109.5\n114.6\n94.5\n104.3\n94.2\n125.3\n79.3\n80.5\n101.5\n100.6\n85.1\n71.3\n88.7\n115.2\nH\n7.0\n6.3\n7.1\n7.9\n5.9\n6.7\n9.4\n8.7\n7.4\n6.2\n6.9\n8.2\n7.9\n7.3\n5.8\n5.4\n13.3\n12.6\n11.0\n9.5\n10.4\n11.9\n9.6\n8.9\n13.9\n13.0\n11.8\n10.4\n10.9\n12.1\nW\n33.2\n29.6\n34.1\n37.5\n39.3\n29.6\n33.2\n43.6\n38.1\n53.0\n60.1\n51.8\n46.3\n38.1\n32.6\n36.6\n43.0\n35.4\n28.4\n26.8\n67.4\n62.2\n53.4\n46.6\nPE MODEL\n68.3\n61.3\n54.6\n49.4\n52.7\n61.3\nH\nVI. FORECAST VERIFICATIONS--MONTHLY MEANS FOR 1975\n.80\n.83\n.77\n.69\n.77\n.85\n.79\n.85\n.80\n.82\n.79\n.81\n.79\n.71\n.77\n.83\n.87\n.82\n.85\n.83\n.78\n.83\n.88\n.84\n.79\n.83\n.79\n.75\n.82\n.85\nR\nA. NMC Grid Area (1,977 Grid Points)\n9.5\n7.4\n5.8\n7.]\n8.2\n8.0\n6.1\n4.9\n6.3\n8.0\n9.9\n9.7\n9.8\n7.5\n9.3\n12.3\n17.1\n18.0\n14.9\n11.9\n14.2\n18.2\n12.5\n12.6\n15.3\n15.5\n13.7\n12.0\n13.4\n16.0\nW\nPERS\n39.6\n54.9\n54.0\n50.3\n38.7\n29.0\n47.0\n43.9\n34.5\n26.5\n35.1\n48.8\n51.2\n38.4\n50.3\n69.8\n90.5\n91.8\n74.4\n57.6\n74.7\n100.0\n65.9\n64.3\n81.7\n78.7\n65.5\n55.5\n69.2\n89.6\n24 hours\nH\n8.5\n7.6\n6.4\n5.9\n6.5\n6.8\n7.3\n6.4\n5.4\n5.1\n5.4\n5.9\n9.0\n6.5\n5.6\n6.1\n7.0\n11.4\n9.9\n8.5\n7.8\n10.7\n9.5\n8.4\n12.1\n11.1\n10.2\n9.3\n9.6\n10.3\nW\n25.2\n28.0\n30.2\n39.0\n32.6\n27.1\n25.0\n28.0\n35.4\n37.2\n30.8\n24.1\n23.8\n46.6\n38.4\n30.8\n26.5\n29.6\n57.3\n49.4\n43.0\n37.5\n41.8\n47.9\nPE MODEL\n58.5\n49.1\n43.6\n40.5\n42.1\n49.4\nH\n.76\n.85\n.77\n.81\n.75\n.63\n.65\n.76\n.84\n.81\n.74\n.82\n.86\n.74\n.79\n.77\n.80\n.85\n.82\n.77\n.83\n.88\n.76\n.82\n.77\n.82\n.78\n.72\n.81\n.85\nR\n1000 mb\n850 mb\n500 mb\n300 mb\nJuly\n200 mb\nSep.\nNov.\nJuly\nJuly\nSep.\nNov.\nJan.\nJuly\nJan.\nSep.\nNov.\nJan.\nMar.\nMar\nJuly\nMay\nSep.\nNov.\nJan.\nMar.\nSep.\nNov.\nMar.\nMay\nJan.\nMay\nMar.\nMay\nMay","19.6\n20.4\n18.2\n16.6\n18.4\n24.6\n23.6\n26.3\n21.8\n16.5\n20.7\n28.4\n17.5\n18.6\n13.8\n10.1\n13.0\n19.2\n10.6\n11.5\n7.9\n6.3\n8.1\n11.0\n12.8\n11.4\n9.7\n7.4\n10.5\n12.8\nW\nPERS\n114.9\n111.6\n95.4\n92.7\n100.6\n155.2\n131.7\n139.9\n111.3\n91.5\n111.3\n169.8\n97.9\n102.7\n74.4\n57.9\n71.6\n117.1\n63.1\n66.8\n45.7\n36.3\n45.4\n68.6\n73.2\n67.4\n51.2\n39.3\n55.8\n76.2\nH\n16.4\n14.8\n11.9\n10.8\n11.5\n15.8\n17.3\n16.3\n12.4\n9.8\n11.8\n17.3\n12.1\n11.3\n8.5\n6.3\n7.4\n11.7\n8.7\n8.4\n6.3\n5.5\n6.1\n8.1\n10.5\n9.6\n8.1\n6.4\n7.5\n9.8\n48 hours\nW\nPE MODEL\n86.0\n70.7\n54.5\n49.7\n57.6\n83.8\n90.9\n79.3\n56.4\n46.6\n56.7\n88.4\n64.9\n57.9\n39.6\n30.5\n35.7\n61.3\n43.6\n41.8\n30.8\n27.1\n29.0\n38.5\n48.5\n45.7\n37.8\n30.5\n33.8\n47.6\nH\n.70\n.83\n.83\n.85\n.84\n.82\n.74\n.84\n.86\n.85\n.86\n.83\n.77\n.84\n.84\n.84\n.87\n.83\n.75\n.82\n.76\n.78\n.83\n.81\n.77\n.79\n.71\n.71\n.81\n.81\nR\n9.9\n18.2\n5.8\n6.9\nW\nBAROTROPIC\n50.9\n65.2\n31.4\n35.1\nH\n.85\n.73\n.78\n.83\nR\n19.5\n19.8\n16.7\n14.8\n16.7\n23.0\n24.0\n26.2\n20.3\n14.8\n19.1\n27.0\n17.7\n18.4\n12.8\n9.1\n12.0\n18.2\n10.9\n11.6\n7.6\n6.1\n7.7\n10.8\n13.4\n11.0\n9.4\n7.3\n9.9\n12.8\n36 hours\nW\nPERS\n108.5\n104.3\n82.3\n79.6\n88.1\n137.5\n128.4\n133.5\n98.5\n79.6\n98.5\n151.8\n94.5\n97.3\n66.2\n50.6\n63.4\n104.0\n61.6\n63.1\n41.5\n33.5\n41.2\n64.0\n73.5\n61.9\n47.3\n36.9\n51.2\n72.6\nH\n13.6\n12.8\n10.6\n8.9\n9.8\n12.7\n14.6\n13.9\n10.8\n8.6\n10.1\n14.1\n10.5\n9.7\n7.2\n5.4\n6.3\n9.8\n7.6\n7.4\n5.5\n4.9\n5.2\n7.1\n9.5\n8.7\n6.9\n6.0\n6.4\n8.5\nW\nPE MODEL\n65.5\n56.7\n43.6\n40.5\n45.4\n62.8\n71.0\n62.5\n44.8\n38.7\n44.8\n67.1\n54.3\n47.3\n32.3\n25.6\n28.4\n48.2\n36.3\n35.4\n26.5\n24.1\n22.9\n35.1\n41.2\n38.1\n30.8\n27.4\n27.1\n39.6\nH\n.80\n.87\n.86\n.86\n.85\n.87\n.82\n.89\n.88\n.86\n.88\n.87\n.81\n.88\n.86\n.85\n.89\n.87\n.81\n.85\n.78\n.74\n.84\n.84\n.82\n.83\n.76\n.71\n.85\n.84\nR\nB, North America--Area 1 (195 Grid Points)\n17.3\n17.2\n13.8\n12.0\n13.2\n19.3\n22.0\n23.0\n16.9\n12.1\n15.3\n23.4\n16.3\n16.0\n10.6\n7.5\n9.5\n15.5\n10.1\n9.9\n6.1\n4.9\n6.2\n9.3\n12.3\n9.7\n7.6\n6.1\n8.1\n11.1\nW\nPERS\n89.9\n83.8\n64.9\n60.4\n65.5\n106.4\n110.4\n110.7\n78.4\n61.0\n75.0\n120.7\n81.4\n79.6\n51.5\n38.7\n47.6\n81.7\n53.5\n51.2\n32.3\n25.9\n31.7\n52.1\n63.7\n50.3\n36.9\n29.6\n39.9\n59.8\n24 hours\nH\n10.8\n10.1\n9.1\n8.2\n8.2\n10.0\n11.6\n10.9\n9.1\n7.5\n8.4\n11.0\n8.7\n7.7\n6.0\n4.9\n5.3\n7.7\n6.5\n5.8\n4.9\n4.5\n4.5\n5.6\n7.9\n7.6\n6.3\n5.3\n5.7\n6.8\nW\nPE MODEL\n47.3\n40.2\n36.3\n34.8\n33.9\n45.7\n54.3\n44.5\n35.7\n31.7\n34.1\n49.7\n43.3\n34.1\n25.0\n22.6\n22.9\n36.3\n30.8\n26.2\n22.6\n21.6\n20.1\n26.8\n33.8\n32.6\n27.1\n24.4\n23.8\n30.5\nH\n.72'\n.84\n.89\n.84\n.83\n.85\n.89\n.86\n.91\n.88\n.85\n.88\n.90\n.83\n.90\n.86\n.81\n.87\n.88\n.83\n.88\n.76\n.69\n.81\n.86\n.85\n.81\n.66\n.82\n.86\nR\n200 mb\n300 mb\n500 mb\n850 mb\n1000 mb\nJuly\nJuly\nJuly\nJan.\nMar.\nSep.\nNov.\nJan.\nMar.\nSep.\nNov.\nJan.\nSep.\nJuly\nJuly\nMar.\nNov.\nJan.\nMar.\nSep.\nNov.\nJan.\nSep.\nNov.\nMar.\nMay\nMay\nMay\nMay\nMay","10.8\n10.0\n7.9\n6.7\n9.5\n10.4\n12.1\n10.9\n8.5\n7.0\n10.2\n11.4\n14.3\n22.5\n24.0\n20.3\n18.9\n22.7\n27.1\n16.4\n17.1\n12.2\n15.3\n18.4\n18.7\n18.8\n15.7\n16.2\n19.8\n22.4\nW\nPERS\n42.1\n69.5\n81.1\n87.2\n71.0\n54.3\n72.7\n71.6\n88.1\n71.0\n102.0\n125.0\n79.6\n67.7\n54.0\n154.6\n124.1\n127.7\n103.4\n88.7\n129.9\n54.0\n141.8\n131.7\n104.9\n145.3\n179.3\n105.5\n110.7\n95.4\nH\n6.4\n7.0\n7.7\n6.4\n5.6\n4.7\n8.9\n7.5\n6.5\n5.4\n7.2\n7.8\n6.9\n9.5\n10.1\n13.7\n11.6\n11.0\n14.5\n14.7\n10.6\n9.7\n8.1\n13.5\n11.4\n10.5\n10.8\n13.0\n13.7\n15.1\n48 hours\nW\n39.0\n44.8\n58.2\n45.1\n37.8\n29.9\n42.1\n50.3\n39.0\n50.3\n40.2\n33.8\n26.5\n58.5\n62.5\n57.0\n80.2\n78.4\n91.8\n79.0\n64.3\n57.9\n85.4\n89.0\n65.9\n57.3\n47.6\nPE MODEL\n84.5\n69.8\n58.5\nH\n.80\n.81\n.81\n.76\n.78\n.75\n.74\n.80\n.81\n.83\n.81\n.86\n.77\n.79\n.78\n.85\n.86\n.82\n.81\n.86\n.80\n.85\n.85\n.75\n.83\n.82\n.78\n.79\n.86\n.78\nR\n9.4\n15.4\n6.6\n8.8\nW\nBAROTROPIC\n57.3\n91.8\n38.4\n57.0\nH\n.82\n.69\n.77\n.80\nR\n11.0\n8.7\n9.6\n9.9\n7.7\n6.4\n9.3\n10.4\n13.0\n9.9\n9.1\n7.2\n6.2\n25.2\n15.7\n15.6\n11.1\n14.4\n17.0\n18.2\n17.4\n14.5\n14.9\n18.9\n20.8\n22.1\n22.0\n18.5\n17.3\n21.8\n36 hours\nW\nPERS\n61.6\n72.3\n76.5\n62.5\n48.5\n38.1\n63.4\n78.7\n161.0\n97.3\n97.9\n82.6\n62.8\n93.9\n112.2\n70.1\n58.8\n47.9\n37.5\n134.8\n138.1\n113.1\n93.9\n134.6\n117.1\n114.3\n89.3\n78.7\n118.6\n139.0\nH\n6.0\n5.5\n5.8\n7.5\n6.4\n5.7\n4.8\n6.3\n7.3\n6.2\n8.0\n8.4\n6.8\n5.6\n5.0\n4.3\n11.1\n10.1\n9.7\n9.7\n14.3\n10.9\n12.7\n11.4\n10.2\n9.6\n12.4\n12.1\n9.2\n8.1\nW\n32.0\n25.6\n34.1\n39.3\n41.5\n32.6\n29.3\n23.5\n32.3\n36.6\n45.1\n36.0\n69.2\n40.5\n34.5\n47.9\n49.1\n68.3\n54.0\n45.7\n74.1\n62.2\n53.7\n49.4\nPE MODEL\n66.2\n56.7\n50.3\n48.2\n63.1\n61.3\nH\n.84\n.85\n.80\n.82\n.79\n.77\n.85\n.84\n.81\n.82\n.77\n.75\n.85\n.82\n.86\n.89\n.82\n.89\n.86\n.83\n.86\n.90\n.83\n.88\n.80\n.86\n.81\n.79\n.85\n.89\nR\n7.4\n7.9\n10.2\n8.5\n6.6\n5.4\n8.4\n8.9\n10.6\n8.9\n7.5\n6.0\n5.0\n8.8\n12.2\n13.8\n20.2\n18.3\n15.2\n14.0\n18.7\n20.8\n20.8\n12.2\n16.3\n14.0\n11.8\n12.3\n15.9\n16.9\nW\nC. Europe--Area 3 (143 Grid Points)\nPERS\n65.2\n39.6\n30.8\n53.7\n61.9\n58.5\n46.3\n37.2\n28.7\n49.4\n54.9\n50.3\n82.0\n75.6\n63.7\n47.6\n74.7\n86.3\n117.4\n107.3\n88.7\n72.3\n110.1\n126.5\n100.0\n85.5\n68.6\n61.9\n95.1\n106.7\n26 hours\nH\n4.5\n4.7\n6.3\n5.2\n4.7\n4.4\n5.1\n5.2\n5.9\n5.1\n6.6\n6.8\n5.7\n4.5\n4.2\n3.8\n10.1\n9.0\n8.2\n7.8\n10.2\n9.7\n7.4\n6.5\n9.4\n8.2\n7.9\n8.2\n9.4\n9.0\nW\n25.3\n22.0\n27.4\n29.9\n25.0\n28.4\n34.8\n27.7\n36.0\n32.9\n26.8\n23.2\n19.8\n31.7\n26.8\n38.1\n39.0\n39.3\n47.9\n47.3\n56.1\n49.1\n41.8\n37.8\n51.5\n52.1\n41.8\n36.9\nPE MODEL\n53.5\n46.6\nH\n.85\n.90\n.85\n.85\n.83\n.83\n.77\n.71\n.87\n.81\n.81\n.78\n.73\n.88\n.85\n.81\n.87\n.80\n.87\n.89\n.86\n.84\n.91\n.85\n.87\n.83\n.85\n.77\n.87\n.89\nR\n1000 mb\n850 mb\n500 mb\n300 mb\nJuly\n200 mb\nJuly\nSep.\nNov.\nJuly\nMar.\nJuly\nSep.\nNov.\nJan.\nSep.\nNov.\nJan.\nMar.\nJuly\nSep.\nNov.\nJan.\nMar.\nMay\nJan.\nMar.\nJan.\nMar.\nSep.\nNov.\nMay\nMay\nMay\nMay","16.6\n19.3\n18.5\n15.8\n16.4\n18.9\n19.0\n21.8\n20.6\n14.4\n16.3\n21.8\n29.0\n14.2\n13.4\n88.7\n10.3\n14.6\n8.1\n9.1\n8.6\n6.2\n7.3\n9.1\n9.6\n11.0\n10.2\n7.3\n8.6\n11.3\nW\nPERS\n100.3\n111.9\n92.1\n77.1\n88.1\n110.7\n112.2\n121.3\n103.4\n68.3\n86.3\n125.3\n83.2\n78.9\n69.5\n43.3\n55.8\n86.9\n48.4\n53.0\n46.6\n32.6\n38.7\n55.5\n55.5\n62.8\n54.3\n37.2\n44.2\n68.0\nH\nOperational 6-layer primitive equation\n14.8\n14.9\n13.5\n12.1\n12.4\n13.1\n12.9\n13.7\n12.5\n10.1\n10.8\n12.6\n9.0\n9.3\n16.8\n6.6\n7.1\n8.4\n7.9\n8.3\n7.0\n5.9\n6.3\n7.8\n10.5\n10.2\n8.4\n7.3\n88.6\n9.2\n48 hours\nW\nPE MODEL\n79.9\n70.4\n66.5\n58.2\n62.5\n65.5\n67.7\n67.7\n61.0\n46.6\n54.0\n41.3\n46.3\n47.0\n39.6\n30.8\n36.6\n42.1\n37.5\n41.8\n32.0\n27.4\n29.6\n35.7\n45.1\n31.4\n45.1\n36.9\n35.4\n40.5\nH\nOperational barotropic model\nbaroclinic forecast model\n.74\n.82\n.77\n.72\n.75\n.83\n.82\n.85\n.82\n.75\n.79\n.87\n.83\n.83\n.82\n.73\n.78\n.86\n.74\n.76\n.77\n.67\n.74\n.81\n.73\n.79\n.77\n.66\n.71\n.83\nR\nPersistence forecast\n8.8\n17.6\n6.2\n7.0\nBAROTROPIC\nW\n56.1\n76.8\n34.1\n41.8\nH\n.63\n.57\n.51\n.59\nR\nBAROTROPIC\nPE MODEL\n16.4\n17.8\n17.7\n14.6\n15.1\n17.3\n18.4\n20.2\n19.3\n13.1\n15.1\n19.9\n13.5\n13.5\n12.4\n7.9\n9.6\n13.4\n7.9\n8.7\n7.8\n5.6\n6.8\n8.6\n36 hours\n9.4\n10.7\n9.7\n6.8\n8.4\n10.8\nW\nPERS\nPERS\n93.0\n98.2\n85.7\n69.2\n78.4\n98.2\n102.4\n107.9\n93.6\n60.4\n77.4\n111.0\n75.0\n71.3\n62.2\n38.1\n50.0\n76.8\n44.5\n48.2\n41.8\n29.3\n34.8\n50.6\n51.5\n58.2\n51.2\n33.8\n41.2\n62.2\nH\n13.5\n13.2\n12.0\n10.7\n10.8\n11.6\n12.1\n12.1\n11.0\n8.9\n9.5\n11.1\n8.7\n8.2\n7.4\n5.9\n6.5\n7.5\n7.6\n7.6\n6.3\n5.6\n5.8\n7.0\n9.8\n9.3\n7.9\n6.9\n7.9\n8.4\nW\nW Root-mean-square vector geostrophic wind error in meters\nPE MODEL\n65.2\n61.0\n60.4\n49.7\n50.0\n50.6\n50.7\n59.1\n52.4\n39.0\n43.9\n48.5\n41.2\n41.8\n34.8\n26.8\n30.5\n34.8\n34.1\n36.6\n28.7\n27.4\n26.8\n31.4\n39.9\n38.7\n36.0\n32.0\n32.9\n36.0\nH\nR Correlation coefficient of forecast and actual\n.79\n.83\n.76\n.73\n.79\n.86\n.84\n.85\n.84\n.78\n.83\n.90\n.84\n.84\n.84\n.74\n.80\n.88\n.75\n.79\n.78\n.61\n.73\n.83\n.74\n.81\n.75\n.58\n.69\n.84\nR\n13.6\n15.0\nH Standard Deviation of Error in meters\n14.9\n12.6\n12.8\n14.7\n15.8\n17.3\n16.4\n11.3\n12.7\n17.0\n11.7\n11.6\n10.5\n6.8\n8.1\n11.4\n7.1\n7.7\n7.0\n4.7\n5.8\n7.3\n8.6\n9.4\n8.5\n5.6\n6.9\n9.1\nW\nPERS\nD. Asia--Area 4 (275 Grid Points)\n72.3\n77.7\n67.4\n56.1\n62.2\n76.8\n82.0\n86.9\n74.7\n49.1\n61.6\n88.1\n60.4\n57.6\n49.1\n30.2\n39.6\n60.7\n37.2\n40.2\n34.1\n22.6\n27.7\n40.9\n43.3\n48.9\n41.2\n25.9\n32.0\n50.0\n24 hours\nH\n12.8\n11.1\n10.5\n9.9\n9.6\n10.2\n10.5\n10.2\n9.6\n8.3\n8.6\n9.5\n7.7\n7.4\n6.6\n5.5\n5.9\n6.5\n7.5\n6.9\n5.8\n5.3\n5.5\n6.5\n9.3\n8.2\n6.9\n6.2\n7.3\n7.1\nW\nPE MODEL\n64.3\n50.9\n46.0\n43.3\n43.3\n44.5\n51.8\n46.3\n43.6\n35.7\n38.1\n41.2\n40.2\n35.4\n31.1\n25.3\n27.1\n31.1\n35.6\n31.4\n26.8\n24.1\n23.8\n29.6\n38.1\n33.2\n29.6\n25.6\n27.1\n29.0\nH\nheight change\nper second\n.71\n.82\n.78\n.69\n.75\n.83\n.81\n.85\n.82\n.71\n.79\n.88\n.80\n.81\n.80\n.66\n.76\n.86\n.67\n.75\n.74\n.57\n.69\n.78\n.69\n.79\n.75\n.58\n.68\n.84\nR\n200 mb\n300 mb\n500 mb\n850 mb\n1000 mb\nJuly\nJuly\nJan.\nMar.\nJuly\nSep.\nNov.\nJuly\nJan.\nMar.\nNov.\nSep.\nJan.\nSep.\nNov.\nJan.\nJuly\nNov.\nMar.\nSep.\nJan.\nMar.\nSep.\nNov.\nMay\nMar\nMay\nMay\nMay\nMay","-38-\nor\n001\n06\nof\nAREA\n4\n001\n06\n08\nos\n09\nO\nNS\n20\nAREA 3\nAREA 5\n+\n50\n60\nAREA\n2\n60\n70\n80\n517\nAREA\n1\nF30\nVERIFICATION\nor.\nAREAS\n570\n80","-39-\nPrecipitation Threat Scores* for the Limited-Area Fine-Mesh Model for 1975\nThreat scores are computed at 0000 and 1200 GMT for a 60-station network\ncovering the 48 states and averaged for each month. These values are\nfor the occurrence of precipitation within the 12-hr period before the\nend of the forecast period. For example, under the column \"36-hr Fore-\ncasts\" the threat scores are based on the forecast and occurrence of\nprecipitation between the 24th and 36th hour.\nTHREAT SCORES\nMonth\n24-hr Forecast\n36-hr Forecast\nJan.\n42.36\nFeb.\n45.24\nMar.\n47.39\nApr.\n43.28\nMay\n35.04\nJune\n34.50\nJuly\n29.57\n24.77\nAug.\n29.12\n26.42\nSep.\n38.10\n33.54\nOct.\n45.58\n38.46\nNov.\n43.17\n37.03\nDec.\n41.35\n32.71\nAnnual\n43.8\n31.8 (6-months)\nH\n* Threat Score\n= F + 0 - H X 100\nwhere\nH = Number of correct forecasts\nF = Number of precipitation forecasts\no = Number of precipitation occurrences.","-40-\nPART VII PUBLICATIONS\nAlaka, M.A., J.P. Charba and R.C. Elvander, Short Range\nThunderstorm Forecasting for Aviation, Dept. of Transport\nReport No. FAA-RD-75, September 1975, 24 pp.\nBergman, K.H., and T.N. Carlson, Objective Analysis of\nAircraft Data in Tropical Cyclones, Mon. Wea. Rev.,\nVo1. 103, No. 5, May 1975, pp. 431-444.\nBermowitz, R.J., An Application of Model Output Statistics\nto Forecasting Quantitative Precipitation, Mon. Wea. Rev.,\n(in press).\nCarter, G.M., Automated Prediction of Surface Wind from\nNumerical Model Output, Mon. Wea. Rev., Vol. 103, October\n1975, pp. 866-873.\nCharba, J.P., Operational Scheme for Short Range Forecasts of\nSevere Local Weather, Proceedings of Ninth Conference on\nSevere Local Storms, April 14, 1975, pp. 51-57.\nDallavalle, J.P., and L.F. Bosart, A Synoptic Investigation\nof Anticyclogenesis Accompanying North American Polar Air\nOutbreaks, Mon. Wea. Rev., Vol. 103, No. 11, November 1975,\npp. 941-957.\nElvander, R.C., The Relationship Between Digital Weather Radar\nData and Reported Severe Weather Occurrences, 16th Radar\nMeteorology Conference, April 22-24, 1975, Houston, Texas,\npp. 333-336.\nFinger, F.G., and A.O. Snow, The Role of Upper Air Observations\nin Mesoscale Networks, AMS Proceedings, February 1975.\nA.J. Miller, D.F. Heath, A.J. Krueger, and\n,\nK. Labitzke, Examples of Synoptic Changes in Ozone and\nCirculation in the Stratosphere, Space Research XVIII,\nVarna, Bulgaria, May 1975.\nM.E. Gelman, F.J. Schmidlin, R. Leviton, and\n,\nB.W. Kennedy, Compatibility of Meteorological Rocketsonde\nData as Indicated by International Comparison Tests,\nJ. Atmos. Sci., Vol. 32, No. 9, September 1975, pp. 1705-1714.","-41-\nFoster, D.S., , and R.M. Reap, Thunderstorm and Severe Local\nStorm Distributions for 1974 Derived from Manually\nDigitized Radar Data and Severe Local Storm Reports,\nProceedings of Ninth Conference on Severe Local Storms,\nApril 14, 1975, pp. 64-67.\nGelman, M.E., and A.J. Miller, Rocketsonde/Rawinsonde\nSatellite Comparisons, Conference Workbook of the Second\nConference on Atmospheric Radiation, October 1975.\nGlahn, H.R., and J.R. Bocchieri, Objective Estimation of the\nConditional Probability of Frozen Precipitation, Mon. Wea.\nRev., Vo1. 103, No. 1, January 1975, pp. 3-15.\nand W.H. Klein, Present Status of Statistical\n,\nWeather Forecasting, Fourth Conference on Probability and\nStatistics, Tallahassee, Florida, November 18, 1975.\nKistler, R.E., and R.D. McPherson, On the Use of Local Wind\nCorrection Technique in Four-Dimensional Data Assimilation,\nMon. Wea. Rev., Vol. 103, No. 5, May 1975, pp. 445-449.\nKlein, W.H., and G.A. Hammons, Maximum/Minimum Temperature\nForecasts Based on Model Output Statistics, Mon. Wea. Rev.,\nVo1. 103, No. 9, September 1975, pp. 796-806.\nLowry, D.A., Meteorological Forecast Applications Associated\nwith AFOS, IEEE Transactions on Geoscience Electronics,\nVol. GE-13, No. 3, July 1975, pp. 116-122.\nMathur, M.B., Development of Banded Structure in a Numerically\nSimulated Hurricane, J. Atmos. Sci., Vol. 32, No. 3, March\n1975, pp. 512-522.\nMcPherson, R.D., Progress, Problems, and Prospects in\nMeteorological Data Assimilation, Bull. AMS.,\nVol. 56, No. 11, November 1975, pp. 1154-1166.\nMiller, A.J., F.G. Finger, and D.F. Heath, Variability of\nStratospheric Ozone Fields Associated with Midwinter\nStratospheric Warmings, I.U.G.G. Proceedings in Grenoble,\nFrance, August 1975.","-42\nMiller, A.J., , and A. Gruber, Statistical Evidence of\nTropospheric Mid-latitude Tropical Interactions, G. R. L.,\nVol. 2, No. 4, April 1975.\nMeteorological Studies on Satellite Ozone Data,\n,\nWorkshop Proceedings of the Stratosphere and Mesosphere:\nDynamics, Physics, and Chemistry Conference (NCAR), July\n1975.\nNagatani, R.M., A.J. Miller, and D.F. Heath, Temporal and\nSpatial Changes in Ozone from the Nimbus IV BUV Experiment,\nProceedings of the International Conference on Environmental\nSensing and Assessment, September 1975.\nQuiroz, R.S., The Stratospheric Evolution of Sudden Warmings\nin 1969-74 Determined from Measured Infrared Radiation\nFields, J. Atmos. Sci., Vol. 32, No. 1, pp. 211-224.\nDynamical Structure Changes of the Stratosphere\nRevealed by SIRS and Related Sounders, Workshop Proceedings\n,\nof the Stratosphere and Mesosphere: Dynamics, Physics, and\nChemistry Conference (NCAR), July 1975.\nA. J. Miller, and R.M. Nagatani, A Comparison of\nObserved and Simulated Properties of Sudden Stratospheric\n,\nWarmings, J. Atmos. Sci., Vol. 32, No. 9, September 1975,\npp. 1723-1736.\nReap, R.M., , and D.S. Foster, New Operational Thunderstorm and\nSevere Storm Probability Forecasts Based on Model Output\nStatistics (MOS), Proceedings of Ninth Conference on\nSevere Local Storms, October 21-23, 1975, pp. 58-63.\nThomas, A.R., , Quality Control of Upper Air Data from a Meso-\nscale Network, AMS Proceedings, February 1975.\nUpper Air Branch Staff, Synoptic Analyses, 5-, 2-, and 0.4-\nmillibar Surfaces for January 1972 through June 1973,\nNASA SP-3091, 1975.","-43-\nNMC OFFICE NOTES\n106 - Note on Insertion Procedures for Meteorological\nData Assimilation\nRonald D. McPherson and Robert E. Kistler\n107 - A Comparison of Radiosonde Reports Received by Four\nMajor U.S. Global Data Processing Centers\nArthur R. Thomas\n108 - NMC Archives\nRobert Gelhard\n109 - The NMC Operational Global Energy Program\nAlvin J. Miller, William Collins, and Dean Dubofsky\n110 - Progress, Problems, and Prospects in Meteorological Data\nAssimilation\nRonald D. McPherson\n111 - Theory of Smoothing Discrete Functions\nFrederick G. Shuman and Frederick G. D. Shuman (U. of Md. )\n112 - (not released)\n113 - (not released)\n114 - Comparisons between NESS and Wisconsin Cloud-Tracked Winds\nPaul Lemar and William Bonner\n115 - The NMC Front End to the NOAA 360/195 System H. A. Bedient\n116 - Multivariate Objective Analysis of Temperature and Wind\nFields\nK. H. Bergman\n117 - Airway Surface Observations (SA's) Stored on the 360/195\nSystem (Hourly Surface Reports)\nWilliam K. Byle and James E. McDonell\n118 - (not released)\n119 - Analysis Error as a Function of Observation Density for\nSatellite Temperature Soundings with Spatially Correlated\nErrors\nKenneth H. Bergman and William D. Bonner\n120 - The Economics of Truncation Error Control Versus High\nResolution\nFrederick G. Shuman","-44-\nTDL OFFICE NOTES\n75-1 - Comparative Verification of Local and Guidance Surface\nWind Forecasts No. 2\nG.M. Carter, H.R. Glahn and G.W. Hollenbaugh\n75-2 - Computer Programs for the MOS Development System\nIBM 360/195 Version\nH.R. Glahn, G.W. Hollenbaugh and F. T. Globokar\n75-3 - Testing the LFM for PoP Forecasting\nH.R. Glahn and J.R. Bocchieri\n75-4 - Automated Prediction of Thunderstorms, Drizzle, Rain, and\nG.M. Carter\nShowers--] No. 2.\n75-5 - Forecasting Surface Wind Direction Using Deviations from\nG.M. Carter\nthe PE Boundary Wind\n75-6 - Evaluation of Computer-Produced Lake Erie Storm Surge\nForecasts from September 1, 1974 through April 30, 1975.\nW.S. Richardson\n75-7 - Comparative Verification of Local and Guidance Cloud Amount\nG.M. Carter\nForecasts No. 1\n75-8 - Simple Properties of Chapeau Functions and Their\nApplication to the Solution of the Advective Equation\nP.E. Long, Jr. and F.J. Hicks\n75-9 - Comparative Verification of Local and Guidance Surface\nWind Forecasts--No. 3 G.M. Carter and G.W. Hollenbaugh\n75-10 - Current Status of Probability of Precipitation Amount (POPA)\nR.J. Bermowitz and E.A. Zurndorfer\nForecasting","-45-\nTECHNICAL MEMORANDA\nERL\nARL-49 Two Case Studies Correlating the Baseline CO2 Record\nat Mauna Loa with Meteorological and Oceanic Parameters\nA.J. Miller, J.M. Miller and R.M. Rotty\nNWS NMC-56\nTheoretical and Experimental Comparison of Selected\nTime Integration Methods Applied to Four-Dimensional\nData Assimilation\nR.D. McPherson and R.E. Kistler\nNWS TDL-54\nClimatology of Lake Erie Storm Surges at Buffalo and\nToledo\nN.A. Pore, H.P. Perrotti and S.W. Richardson\nNWS TDL-55\nSome Physical and Numerical Aspects of Boundary Layer\nModeling at the Techniques Development Laboratory\nP.E. Long and W.A. Shaffer\nNWS TDL-56 Dissipation, Dispersion, and Difference Schemes\nP.E. Long\nNWS TDL-57 A Predictive Boundary Layer Model\nW.A. Shaffer and P.E. Long\nNWS TDL-58 A Preliminary View of Storm Surges Before and After\nStorm Modifications for Alongshore Moving Storms\nC.P. Jelesnianski and C.S. Barrientos","-46-\nABBREVIATIONS AND ACRONYMS\nAutomation of Field Operations and Services\nAFOS\nAmerican Meteorological Society\nAMS\nBoundary Layer Model\nBLM\nCourant-Frederichs-Lewy\nCFL\nCentral Processing Unit\nCPU\nData Systems Test\nDST\nFirst Garp Global Experiment\nFGGE\nGlobal Atmospheric Research Program\nGARP\nGeophysical Fluid Dynamics Laboratory\nGFDL\nGoddard Institute for Space Studies\nGISS\nHigh-resolution Infrared Sounder\nHIRS\nInternational Business Machines\nIBM\nInfrared Temperature Profile Radiometer\nITPR\nKeyboard Cathode Ray Tube\nKCRT\nLimited-area Fine-mesh Model\nLFM\nLimb Radiance Infrared Radiometer\nLRIR\nManually Digitized Radar\nMDR\nModel Output Statistics\nMOS\nMeteorological Satellite Laboratory\nMSL\nMonthly Weather Review\nMWR\nNational Aeronautics and Space Administration\nNASA\nNational Center for Atmospheric Research\nNCAR\nNational Environmental Satellite Service\nNESS\nNational Meteorological Center\nNMC\nNational Oceanic and Atmospheric Administration\nNOAA\nNumerical Weather Prediction\nNWP\nNational Weather Service\nNWS\nNational Severe Storms Forecast Center\nNSSFC\nPrimitive Equation\nPE\nProbability of Frozen Precipitation\nPoF\nPressure Modulated Radiometer\nPMR","-47-\nQPF\nQuantitative Precipitation Forecasts\nRADAT\nRadiosonde Early Transmission\nREEP\nRegression Estimation of Event Probability\nRMS\nRoot-mean-square\nSPLASH\nSpecial Programs to List Amplitudes of Surges\nof Hurricanes\nSST\nSupersonic Transport\nSTRATWARM\nStratospheric Warmings\nTAP\nTerminal Alerting Procedure\nTWERLE\nTropical Wind, Energy Conversion and\nReference Level Experiment\nUCLA\nUniversity of California at Los Angeles\nVIL\nVertically Integrated Liquid Water Content\nVFM\nVery Fine Mesh Model\nVTPR\nVertical Temperature Profile Radiometer\n3L GLOBAL\nThree-layer Global Model\n8L GLOBAL\nEight-layer Global Model\n8L HEM\nEight-layer Hemispheric Model\n6L PE\nSix-layer Primitive-Equation Model"]}