• Cover


    J.C. Sadler, M.A. Lander, A.M. Hori, and L.K. Oda
    Department of Meteorology, University of Hawaii
    UHMET-87-02, March 1987
    (web and CD versions March 2003)


    Over the past 15 years the increasing interest in climate analysis and forecasting has been evidenced by many national and international programs dealing with the problem or using to justify related research¾the North Pacific Experiment (NORPAX); Equatorial Pacific Ocean Climate Studies (EPOCS); Pacific Equatorial Dynamics (PEQUOD); Tropical Pacific Ocean Heat and Mass Budgets (Tropic Heat); Seasonal Equatorial Atlantic Experiment (SEQUAL); Tropical Ocean Global Atmosphere (TOGA); and World Climate Research Program (WCRP). In part, all of these set out to improve the atmospheric and oceanic data bases. EPOCS conceived and partly funded a large joint effort to compile global ship observations by the Environmental Research Laboratories of NOAA; the Cooperative Institute for Research in Environmental Sciences (CIRES) of NOAA and the University of Colorado; the National Center for Atmospheric Research (NCAR); and the National Climatic Data Center (NCDC) of NOAA. The result was the Comprehensive Ocean-Atmospheric Data Set (COADS) for the period 1850-1979. Details of how the ship observations were collected, evaluated, and compiled on a 2° by 2° grid are contained in Woodruff et al. (1987). We selected the last 80 years, 1900-1979, for this atlas.


    The numbers of observations by oceans are shown in Fig. 1. The general increase with time was disrupted during World Wars I and II, reversed in the early 1960's and more or less stabilized in the 1970's at about 155,000 observations per month. Unfortunately the observations are not evenly distributed in either time or space. The number density is three times greater in the Atlantic than in the Pacific or Indian Ocean; within ocean basins the observations are also poorly distributed (Fig. 2). Observations are relatively numerous and, although highly skewed toward the coasts, the coverage is quite reasonable in the well-traveled North Atlantic, to a lesser degree in the Pacific north of 20N, and in the north Indian Ocean. South of 20N the data are concentrated along narrow, shipping routes leaving large poorly covered areas, particularly in the Pacific where 2° grid squares have an average of less than one observation per month. For a few grid squares off the east coast of the United States the number of observations for the 80-year monthly mean exceeds 12,000 while the minimum is less than ten for some squares in the equatorial Pacific near the dateline, many in the southeastern Pacific and some in the central South Atlantic. Ratios of ten to one in number of observations in adjacent grid squares are not uncommon. We have dwelt on the data distribution to emphasize that, although COADS is the best data base yet assembled, there remain large areas of the tropics and Southern Hemisphere with inadequate data for a unique analysis.

    Figure 1. Year-month number in thousands of ship observations per ocean basin.

    Figure 2. The average distribution of ship observations per 2° square per individual month for the 80-year period ;1900-1979. Black- >30; grey = >10<30; chicken tracks = >2<10; no shading = <2; dashed lines enclose areas of <1.


    The inadequate number and uneven distribution of data coupled with large gradients and the relatively small atmospheric systems in the tropics require a manual analysis for optimum results. In this atlas, manual analysis, in contrast to machine analysis, permits a variable radius of data influence (a variable amount of smoothing) and the incorporation of auxiliary information and experience. For example the wind analysis utilizes knowledge of the distribution of sea level pressure, sea surface pressure, sea surface temperature, rainfall, satellite cloudiness; orography and coastal effects; persistent wind systems and patterns; and climatological constraints. Recent atlases, produced by machine methods from essentially the same data base, differ significantly from our atlas (Hastenrath and Lamb, 1979; Reynolds, 1985). When compared to recent wind stress produced by a machine analysis scheme (Picaut et al., 1985), differences amount to 25% in the maxima, minima, and gradients of monthly averaged wind stress.


    Of the 19 available parameters in the COADS, wind, wind stress, sea level pressure and sea surface temperature were selected for publication since they are of interest to the greatest number of people in the World Climate Research Program (WCRP). The four fields are shown [in the original print version] on facing pages for easy intercomparison. Total cloudiness determined from nine years of satellite observations has been published (Sadler et al., 1984). COADS total cloudiness shows identical patterns but poorer data coverage so is not included in the atlas. So as to maintain a comparable reproduction scale for the three oceans, the atlas is in two volumes. Volume I contains the Atlantic and Indian Oceans and Volume II the Pacific Ocean. The data plots for each element contain (1) the number of observations during the 80-year/month period within each 2° square and (2) the 80-year/month average value of the element after weighting each year/month value by number of observations. The observed resultant directions of the wind and wind stress are shown by arrows with the length of shaft proportional to the resultant speed of the wind and value of the stress.

  • Chart 1. Surface Wind The resultant direction is depicted by streamlines and the resultant speed in m/sec by dashed lines. The number of observations plotted over land gives some indication of the number of misplaced observations scattered throughout the data.
  • Chart 2. Sea Level Pressure The isobars are labeled in hectopascals (hPa) minus a thousand. Where needed one-half hPa intervals are shown as dashed lines. The analyses over land utilize long-term mean data based on a 5° grid obtained from the National Center for Atmospheric Research (NCAR).
  • Chart 3. Surface Wind Stress The isoplets represent the magnitude of the wind contribution to the stress in m/sec . To obtain a dimensionally correct value of the wind stress, the numbers should be multiplied by the air density and drag coefficient, which may or may not be constant depending on the empirical formula selected. The analyzed stress direction is assumed to be the same as the resultant wind direction of Chart 1 following the results of Thompson et al. (1983).
  • Chart 4. Sea Surface Temperature The isotherms are labeled in degrees Celsius. Where needed, one-half degree intervals are shown as dashed lines.


    From the analyses, data were manually digitized and recorded for 2( intersections of latitude and longitude. After being keyed onto a computer disk the date were replotted for error checking through inspection and reanalysis. The wind and wind stress data were converted to u and v components and meshed with the sea level pressure and sea surface temperature onto an archival tape. [The maps are printable at 11 x 17; and the values are readable on screen at ~600% enlargement.]
    January Wind
    January Pressure
    January Stress
    January Temperature
    February Wind
    February Pressure
    February Stress
    February Temperature
    March Wind
    March Pressure
    March Stress
    March Temperature
    April Wind
    April Pressure
    April Stress
    April Temperature
    May Wind
    May Pressure
    May Stress
    May Temperature
    June Wind
    June Pressure
    June Stress
    June Temperature
    July Wind
    July Pressure
    July Stress
    July Temperature
    August Wind
    August Pressure
    August Stress
    August Temperature
    September Wind
    September Pressure
    September Stress
    September Temperature
    October Wind
    October Pressure
    October Stress
    October Temperature
    November Wind
    November Pressure
    November Stress
    November Temperature
    December Wind
    December Pressure
    December Stress
    December Temperature


    The analyses of the COADS were supported by the National Oceanic and Atmospheric Administration (NOAA) as part of the Tropical Oceans Global Atmosphere (TOGA) program. Publication of the atlas was supported by the NOAA Equatorial Pacific Ocean Climate Studies (EPOCS) program. This work was supported by NOAA Cooperative Agreement NA-85-ABH-00032.


    • Hastenrath, S., and P. J. Lamb, 1979: Climatic Atlas of the Indian Ocean. The University of Wisconsin Press. Picaut, J., J. Servain, P. Lecomte, M. Seva, S. Lukas, and G. Rougier, 1985: Climatic Atlas of the Tropical Atlantic. Univ. De Bretagne Occidentale and Univ. of Hawaii, 467 pp.
    • Picaut, J., J. Servain, P. Lecomte, M. Seva, S. Lukas, and G. Rougier, 1985: Climatic Atlas of the Tropical Atlantic. Univ. De Bretagne Occidentale and Univ. of Hawaii, 467 pp.
    • Reynolds, R., 1985: Global Mean Monthly Surface Winds. Climate Analysis Center, National Meteorological Center. Personal communication. Picaut, J., J. Servain, P. Lecomte, M. Seva, S. Lukas, and G. Rougier, 1985: Climatic Atlas of the Tropical Atlantic. Univ. De Bretagne Occidentale and Univ. of Hawaii, 467 pp.
    • Sadler, J. C., B. Kilonsky, L. Oda, and A. Hori, 1984: Mean Cloudiness over the Global Tropics from Satellite Observations. Dept. of Meteor., Univ. of Hawaii Report No. UHMET 84-01 and U.S. Navy NEPRF Contractor Report No. CR-84-09, 73 pp. Picaut, J., J. Servain, P. Lecomte, M. Seva, S. Lukas, and G. Rougier, 1985: Climatic Atlas of the Tropical Atlantic. Univ. De Bretagne Occidentale and Univ. of Hawaii, 467 pp.
    • Thompson, K. R., R. F. Marsden, and D. G. Wright, 1983: Estimation of LowFrequency Wind Stress Fluctuations over the Open Ocean. J. Phys. Oceanogr, 13, 1003-1011. Picaut, J., J. Servain, P. Lecomte, M. Seva, S. Lukas, and G. Rougier, 1985: Climatic Atlas of the Tropical Atlantic. Univ. De Bretagne Occidentale and Univ. of Hawaii, 467 pp.
    • Woodruff, S. D., R. J. Slutz, R. L. Jenne, and P. M. Steurer, 1987: A Comprehensive Ocean-Atmosphere Data Set. Bull. Amer. Meteor. Soc., 68, 1239-1239.

    Note: The analyzed grid point data are available on tape from the Department of Meteorology, University of Hawaii, Honolulu, HI 96822, and from the Data Support Section, NCAR, Boulder, CO 80307.

    This page was created 25 March 2003.
    Please send all comments to dhenders@hawaii.edu .

    Thank you!