75 .^'' E&\-^V $)\- -^ NOAA Data Report ERL GFDL-2 S *TES O* * FGGE LEVEL III-B DAILY GLOBAL ANALYSES PART II (MAR 1979 - MAY 1979) Geophysical Fluid Dynamics Laboratory Princeton, New Jersey July 1983 noaa NATIONAL OCEANIC AND ATMOSPHERIC ADMINISTRATION / Environmental Research Laboratories Digitized by the Internet Archive in 2013 http://archive.org/details/fggeleveliiibdaiOOplos 'Mem of NOAA Data Report ERL GFDL-2 FGGE LEVEL III-B DAILY GLOBAL ANALYSES PART II (MAR 1979 - MAY 1979) Jeffrey J. Ploshay Robert K. White Kikuro Miyakoda Geophysical Fluid Dynamics Laboratory Princeton, New Jersey July 1983 UNITED STATES NATIONAL OCEANIC AND Environmental Research DEPARTMENT OF COMMERCE ATMOSPHERIC ADMINISTRATION Laboratories Malcolm Baldrige, John V.Byrne, Vernon E. Derr Secretary Administrator Acting Director U. S. Depository Co*y NOTICE Mention of a commercial company or product does not constitute an endorsement by NOAA Environmental Research Laboratories. Use for publicity or advertising purposes of information from this publication concerning proprietary products or the tests of such products is not authorized. n CONTENTS I. Introduction II. Outline of Analysis Process III. FGGE Ill-b Data Set Format and Contents IV . References V . Acknowl edgements Page . 1 . 1 . 5 . 7 m FGGE LEVEL III-B Daily Global Analyses Part II (Mar 1979 - May 1979) Jeffrey J. Ploshay, Robert K. White and K. Miyakoda I. Introduction This four-part booklet contains daily qlobal analyses of the FGGE Level Ill-b data set produced at GFDL (Geophysical Fluid Dynamics Laboratory). GFDL was one of the two level Ill-b data producers for the entire FGGE year - the other producer was the ECMWF (European Centre for Medium-Range Weather Forecasts). FGGE provided an unprecedented volume of meteorological data, from various sources, for the year 1979. The purposes of FGGE are to observe comprehensively and accurately the atmospheric state and circulation on a global scale, and to increase understanding of the dynamics of the equatorial atmosphere and ocean to ultimately improve weather and climate prediction. For the sake of comparison by prospective level III -b data users, we decided to adopt a format for our booklet similar to the ECMWF booklet already published. II. Outline of Analysis Process The analysis process (4DA) used at GFDL to produce the FGGE Ill-b analyses, similar to Miyakoda et al. (1976), can be called a four- dimensional, continuous, data assimilation system. The system consists of three phases: preprocessing of input (level Il-b) data, dynamic assimilation and initialization. i) Preprocessing Level Il-b data from various sources 9 blocked into 6 hour inter- vals, are sorted by parameter. The parameters are: horizontal wind components u and v, temperature T, geoptential height Z, moisture (humidity r, depression of dewpoint temperature T-T., or mixing ratio of water vapor q), sea level pressure P SL , sea level temperature TV. and surface wind components U $L and V SL . During this step quality control is performed with respect to format, gross error and hydrostatic checks. Next, the data is prepared for insertion in the dynamic analysis. First vertical and then lateral interpolation are performed to obtain gridded information from observation points. Vertical optimum inter- polation (OPI), from arbitrary pressure levels onto 19 mandatory pressure levels is done for all variables. The range in vertical interpolation is limited to 3 consecutive levels. The observed data is checked for its validity by comparing it with the analyzed fields that were produced by 4DA at the most recent synoptic time (I.e., 00 or 12 GMT). After the vertical interpolation, a "buddy" check (a comparative test with neighboring observations) is made at mandatory levels (Bergman, 1978). The lateral interpolation to the model's grid is made by a hori- zontal "local" OPI (Gandin, 1963) from observation locations at the mandatory levels every 2 hours (i.e., 02, 04, 06, 08, 10, 12) for a 12- hour period. Data from two hour time blocks (e.g. 23 to 01 0MT) are collected from a 250 km radius around each grid point, up to a maximum of 8 observations. If there are no observations within range of a gridpoint, no insertion data is produced. The first guess used in the OPI at each grid point is the most recent synoptic time (00 or 12 GMT) 4DA. Weights are also determined for each grid point value based on the reliability (distance and observational error characteristics) of the contributinq observations. The final process in determining insertion data at a grid point is a time interpolation, to 2 hour periods, of values created by the local OPI and the model, in order to get time continuity of the insertion data. The time interpolation scheme involves a hierarchy of four interpolation values, where lower precedence data is used only if a higher precedence value has not yet been obtained. The order of prece- dence is: first the input data from optimum interpolation at the time (if it exists), second the grid points created by the 6 hour inter- polation, third the grid points created by the 12 hour interpolation and, fourth the results of the model simulation of the initial time (00 or 12 GMT). i i ) Dynamic Assimilation The general circulation model used for assimilation is the "spectral transform" model (Gordon and Stern, 1982; Fels and Schwarzkopf, 1975) with an assimilation system developed by Simmonds (1976, 1978) (see Figure 1). The insertion data at mandatory isobaric levels are interpolated to sigma-levels, using cubic splines on a log-p coordinate system. In the insertion process, these data replace the model's solution completely, or partially, depending on the weights assigned to them. If there is no data at a grid point the model's solution is retained. By repeated insertion of the same data at each time step over a 2-hour interval, a balanced state may be achieved that is faithful to the data, and is also consistent with the model's dynamics and physics. The model should impose its own constraints between the variables. Before the data insertion, a quality control check is made by comparing the insertion data with the model's solution. If necessary, the insertion data is modified so that the difference between the model GFDL-DATA ASSIMILATION SYSTEM SUMMARY INSERTION DATA * method * variables * grid * first guess * application ASSIMILATION * method * variables * resolution * time integration * lateral diffusion * vertical diffusion * boundary layer * topography * radiation prepared by 3-dimensional, univariate, optimum interpolation PgL,u,v,T,q on pressure levels (no data inserted at upper two model levels) N40 gaussian, 19 pressure levels 12-hour assimilation results every 12 hrs. (+ 1 hour data window) Model levels a .0022 - 019 052 099 156 223 297 376 458 542 624 .703 777 844 901 948 980 998 L 18 WMWWW* 5cm \ 50 -Land- 5 500 Z tie (km) -1 40. •2 27 -3 20 -4 16 -5 13. ■6 11 -7 9 -8 76 -9 6 2 -10 50 -11 39 -12 30 -13 21 -14 1 4 -15 086 -16 45 -17 017 -18 0.02 GFDL Archiv( levels (mb) 04 2 5 10 100 150 200 250 400 500 700 Sea ff * sea surface temperature * land surface temperature * moisture INITIALIZATION * method * application ARCHIVE * variables * grid data injected into a global spectral model, using weighted time inter- polation P*,T, C (vorticity) , D (divergence) , q on sigma levels R30L18 (rhomboidal truncation at 30 waves, 18 a-levels) semi-implicit (At=8%20 minutes- function of CFL criterion) KV2 mixing length method (to 3 km depth), dry convective adjustment (above level 18) Monin-Obukhov process spectrally truncated developed by Fels and Schwarzkopf (i) clouds - climatological monthly mean for each latitude (ii)application- diurnal variation; short- and long-wave radiation calculation every 2 hours RAND monthly climatological normals, yet varying daily determined by surface heat balance, using 3 soil levels to model heat flux large scale condensation at 80% humidity saturation, cumulus parameterization by moist convective adjustment. non-linear normal mode, 7 vertical modes (only modes with periods shorter than 6 hours adjusted) every 6 hours u,v (m/s), T(°K), mixing ratio (g/g), geopotential height (geop. meters), RH(%) , vertical velocity (mb/sec), surface wind stress - T x , T y (N/ m ) 1.875° latitude/longitude, 19 pressure levels (9 pressure levels for MR and RH) > Stratosphere > Troposphere ^Boundary Layer • Subsurface Layer Figure 1 solution and the insertion data does not exceed a tolerance criterion. Data was not inserted into the model's top two levels, iii) Initialization Every 6 hours a non-linear normal model initialization procedure (developed by Machenhauer, 1977, and Ballish, 1980) is applied to the model solution. It has been designed to control the growth of spurious gravity modes but not to alter the model balance, especially in the tropics. The first 7 of 18 vertical modes are initialized allowing 4 iterations for convergence. In order to not risk initializing out gravity modes that might be important, especially in the tropics, only those modes with periods shorter than 6 hours are adjusted. III. FGGE Ill-b data set format and contents The parameters available in the GFDL/FGGE data set are sea level pressure, horizontal wind components, temperature, mixing ratio of water vapor, geopotential height, relative humidity of water vapor, vertical sigma velocity and surface horizontal wind stress components. Analyses are available, on a regular latitude/longitude grid of 1.875° resolution in both dimensions, at 19 pressure levels (interpolated from model sigma-levels) with relative humidity and mixing ratio only available up to 300 mb. The archive format is the FGGE level Ill-b format (GARP, 1978). The two world data centers through which the data can be obtained are: World Data Center - A for Meteorology National Climatic Center Federal Building Asheville, North Carolina 28801, USA or World Data Center - B for Meteorology All Union Research Inst, for Hydrometeorological Information 6 Koroles Street Obninsk, Kaluga District U.S.S.R. 249020 This booklet contains analyses for March 1979 through May 1979. Each day at OOGMT, four levels of polar stereographic maps of the northern and southern hemispheres (geopotential height at 500 mb, 300 mb, 500 mb, and sea level pressure), as well as two levels (200 mb and 850 mb) of tropical streamlines and isotachs, are presented. The geo- potential maps have a contour interval of 8 decameters and the sea level pressure maps an interval of 5 mb with areas of model topography greater than one kilometer shaded. The streamline maps have isotachs with a varying stippling density as described below: 200 mb 850 mb no shading 0-15 m/s no shading 0-5 m/s light shading 15-30 m/s light shading 5-10 m/s medium shading 30-40 m/s medium shading 10-15 m/s dark shading 45 m/s dark shading 15 m/s IV. References Ballish, B., 1980: Initialization, Theory and Anpli cation to the NMC Spectral Model. Ph. D Thesis, Dept. of Meteorology, University of Maryland, 151 pp. Bergman, K.H., 1978: Role of observational errors in optimum inter- polation analysis. Bull . A mer . Meteor . Soc . , 59, 1603-1611. Fels, S.B. and M.D. Schwarzkopf, 1975: The simplified exchange approxi- mations: a new method for radiative transfer calculations. J_. Atmos . Sci . , 32_, 1475-1488. GARP, 1978: Appendix 11, FGGE Data Management Plan. GARP Implementation Operations Plan, Vol. 3, WMO, Geneva. Gandin, L., 1963: Objective analysis of meteorological fields . Gidro- meteosologicheskoe Izdatel 'stvo, Leningrad, U.S.S.R., 286 pp. Engl. Transl. 1967, U.S. Dent of Commerce, 242 pp. Library of Congress, Washington, DC (NTIS N6618047 0C996 G3313). Gordon, C.T. and W.F. Stern, 1982: A description of the GFDL global spectral model. Mon . Wea . Rev ., 110 , 625-644. Machenhauer, B., 1977: On the dynamics of gravity oscillations in a shallow water model with application to normal mode initialization. G eitr&ge zur Physik der Atmosphare , 50 , 253-271 . Miyakoda, K. , L. Umscheid, D.H. Lee, J. Sirutis, R. Lusen and F. Pratte, 1976: The near-real-time, global, four-dimensional analysts experi- ment during the GATE period, Part I. J_. Atmos . Sci . , 33^ 561-591. Simmonds, I., 1976: Data assimilation with a one-level primitive equation spectral model. J_. Atmos . Sci . , 33 , 1155-1171. Simmonds, I., 1978: The application of a multi-level spectral model to data assimilation. J. Atmos. Sci., 35, 1321-1339. Acknowledgements The authors wish to thank Dr. Joseph Smagorinsky, the former Director of GFDL for his enthusiastic support and guidance and for providing the facilities and environment where this long-term large-scale international effort could be conducted. This type of work can only be possible with the contributions from a large number of people. We appreciate the: valuable assistance provided by Dr. Isidoro Orlanski (acting Director of GFDL), T. Terpstra, F. Uveges and the computer operations staff, B. Wyman, G. Vandenberghe, J. Conner, P. Tunison, W. Ellis, M. Zadworney, B. Williams, P. Baker and R. Lusen. scientific guidance and suggestions from W. Stern, C.T. Gordon, J. Mahlman, J. Sirutis, I. Simmonds, K. Bergman, S. Fels, F. Mesinger, A. Oort, N. Lau, D. Schwartzkopf , R. Caverly and L. Umscheid. technical advice and help by J. Brown, B. Ballish, F. Baer, K. Puri, J. Stackpole, A. Kasahara, K. Labitske, L. Bengtsson, N. Phillips, A. Hollingsworth, D. Parish, R. McPherson, A. Desmaris, R. Jenne; and useful information provided by R. Fleming, T. Schlatter, W.L. Smith, W. Shenk, L. Hubert, D. Gauntlett, P. KSllberg, T.N. Krishnamurti , G. Meleshko, J. Perry, W. McGovern, T. Kanishige, J. Harrison, and H. Lyne. MARCH 1979 10 z o o cc o O t*P ^T r\i (- 1A1 1 „ LO o in •T UJ "T 11 02 MAR 7S ^ 00 GMT 12 GFDL 13 14 15 50 MB Z 16 o o a. O O L^ Z O o (X o O C£ n r\i Y- fM "3" „ 1 (M h iAJ M- o in ITl Q U ' Ld "3 ^r UJ " 17 18 04 MAR 73 19 Z O o cr o O <-£ „ T rsi h- rM "tf u ir< <3 u n 1 u, -T 20 21 05 MAR 79 00 GMT 22 23 Ob MAR 78 00 GMT 24 25 Z o O en O o I* <3- r\i (— "'J M- LP C3 U i T UJ "T 26 50 MB Z 27 28 29 30 50 MB Z 31 32 33 34 z O o cc o o itf- Z _ 3- LP a in n Ld t z o O cr o O tS „ ^r iAl h- ! "-J T . in o in <3 LO in T Ul 1 <3 83 GFDL 84 85 86 87 88 2 o O CL o o u*- z o _ *T I'M t- r\ ^ . ■ ,? 203 z o o cr "3- cm h- in o i UJ o o 204 05 MAY 1°> 00 GMT 50 MB 205 206 207 208 Ob MAY 73 00 GFDL 209 Z O LI i •T a o o US- z o O CC O O f\ <5 *■ r\ h- LP a T UJ in 252 21 MAY 79 00 GMT GFDL 253 254 z o o cr o O U* 1 Z O O ^r fM i- fM "xf _