[Next] [Prev] [Contents] [Top]

An Introduction to Atmospheric and Oceanographic Datasets


There is a broad range of oceanographic data types. This chapter will focus on the data that defines the physical characteristics of the ocean-atmosphere boundary layer and the subsurface distribution of sea water properties. Upper atmosphere observations taken from ocean islands or ships can be considered conventional meteorological observations and are discussed in Chapter 3.

Oceanographic data are collected using both in situ methods and remote sensing. The most obvious remote sensing platforms are satellites, but scientific aircraft, some special buoys, and even some ships use instruments (e.g., radiometers) to remotely sample the ocean surface. Some of the primary remote sensing instruments and resulting oceanographic data are: radiometers which estimate sea surface temperatures (SSTs), scatterometers which measure wave disturbances and yield surface wind speeds and directions, and high precision altimeters that measure ocean surface deformation. The surface deformation is used to estimate sea surface slopes and ocean currents. Satellite data are a major asset for oceanographic research. In situ sampling from ships and buoys does not, in general, yield sufficient spatial or temporal data resolution over the vast ocean regions that cover more than 70% of the planet. Carefully calibrated and adjusted satellite data, sometimes blended with in situ observations as a data processing procedure, provide our best assessment of global ocean conditions.

Useful in situ ocean observations come from different sources, with varying degrees of quality. The highest quality data are collected during scientific research programs, by instrumented buoys (both moored and free drifting), by ships specifically designed to collect environmental data, and by coastal or island stations that function in a manner similar to standard land stations. Lower quality data, but nevertheless quite valuable, are regularly collected aboard merchant ships as they traverse shipping routes, and by fishing fleet vessels during commercial fishing operations.

Scientific research programs collect the widest variety of in situ data. Typical ship board activities will collect sea surface data (SST, salinity, wave height, wave direction, etc.), near-surface meteorological conditions (air temperature, wind speed, wind direction, dew point temperature, barometric pressure, cloudiness, etc.) and, often, subsurface sea water characteristics (e.g., vertical profiles of temperature, salinity, dissolved nutrients, dissolved gases, anthropogenic tracers, ocean currents, and ocean bottom depth). Some research programs also deploy surface drifting buoys whose locations are monitored by satellite. These provide buoy trajectories (that approximate surface ocean circulation), and usually a few other geophysical variables (e.g., SST, barometric pressure, etc.). To a lesser extent, some free drifting buoys are located below the ocean surface. These buoys are tracked acoustically or they periodically rise to the surface for satellite tracking. Buoys of this type are used to monitor subsurface oceanic flow as well as subsurface sea water properties. Moored surface buoys with subsurface instruments below are also used by science programs. The surface instrumentation collects many types of data relevant to ocean-atmosphere boundary layer processes, while the subsurface instruments normally focus on water temperature, salinity, pressure, and ocean currents.

The sampling periods for scientific research programs are often short relative to the needs of climate studies. Some long term environmental monitoring is carried out by NOAA's National Data Buoy Center (NDBC). The NDBC maintains an array of approximately 50 moored surface buoys primarily located in the Atlantic and Pacific coastal waters of the United States, but also include areas off the Hawaiian Islands and Alaska, in the Gulf of Mexico and the Great Lakes. The deployment of this array began in the mid-1970's and is ongoing. Essentially, these buoys represent a quasi-permanent array of ocean observing stations. Weather observing ships were used beginning in 1945. These ships remained positioned at assigned locations serving as weather and ocean/atmosphere observing stations. However, as automated instrumentation and buoy technology improved the deployment of weather ships has dramatically decreased. Although the weather ships did some profile sampling of subsurface ocean properties, the primary focus was on the ocean surface and meteorological conditions. The NDBC buoy array also samples surface and near-surface conditions such as SST, wind speed and direction, air temperature, barometric pressure, and wave data.

One scientific program that is contributing important data for climate studies is called TOGA/Tropical Atmosphere Ocean (TAO). A smaller program began about 1980 and has evolved and expanded into the present day TOGA/TAO which covers the equatorial belt of the Pacific Ocean with surface and subsurface instrumentation fixed to 30-50 moored buoys (Fig. 4.1). In near real time mode, data from these instruments are collected by satellite and are used for global weather and ocean condition forecasts. During periodic instrument service and repair, these data are also collected in a delayed mode. Following post sampling calibration of the instruments, the delayed mode data are quality checked and corrected. Both the near real time and delayed mode data are important for monitoring typical ocean conditions such as El Nino.

Data collected on merchant and fishing vessels are a large source for surface oceanographic data. Typically, these data are gathered at synoptic weather observing times aboard ships that are in transit. Mariners have done this through history, of course, with widely varying methods and degrees of accuracy. The typical measurements are wind speed and direction, barometric pressure, air temperature, SST, and local weather conditions. In the earliest times, these data were recorded by hand in logbooks. Some historical logbook data have been digitized and now the earliest digital records are from the early 1800's. Modern vessels use automated systems whereby the data are collected digitally and transmitted via satellite to land-based collection agencies. These data provide critical information for present day weather and ocean-condition forecasting. Nevertheless, significant amounts of data are still only recorded in logbooks. Several data archival programs are in progress to digitize more logbook data. Given the vast regions of the ocean, and the relatively sparse sampling that occurs, almost any available data are considered useful.

There are many ancillary oceanographic data types that are important. Sea level, sea ice concentration, and ocean bottom topography are a few examples. These are briefly discussed at the end of this chapter.

Table 4.1 shows a list of selected oceanographic datasets that are available at NCAR. This selection of data will be used as example in the subsequent discussion. This is not a comprehensive list of all available data. Many federal agencies and other research institutes have important data not shown in Table 4.1. Some organizations which archive oceanographic data are listed in Appendix A.

Ocean-Atmosphere Boundary Layer Data

Data that define the physical conditions at the ocean-atmosphere boundary are of interest to many scientists. Atmospheric-oceanic forcing is studied over a broad range of temporal and spatial scales. Generally, temporal periods span seconds to centuries and spatial coverages range from centimeters to thousands of kilometers. The existing data are generally inadequate to address a broad range of research topics. On a global scale, the data are marginally adequate for problems focused on long term monthly variations. As necessary spatial and temporal requirements of a particular scientific investigation decrease, it becomes more likely that data will exist to study the problem.

Efforts have been made to collect datasets suitable for global studies of ocean-atmosphere climate. One such project has resulted in the Comprehensive Ocean Atmosphere Data Set (COADS). COADS is a cooperative effort between four groups; NCAR, the Environmental Research Laboratory (ERL) of NOAA, the National Climatic Data Center (NCDC), and the Cooperative Institute for Research in Environmental Sciences (CIRES). The goal of the COADS project is to collect near-surface oceanographic data from many sources and merge them together to form the most complete data base possible. The primary sources included in COADS are: surface observations from merchant ships operated by many different countries, observations taken from oceanographic research vessels and some fishing fleet data, data from moored ocean buoys and coastal stations operated by the NDBC, moored buoy data from the TOGA/TAO experiment, the global drifting buoy data from the Marine Environmental Data Service (MEDS) of Canada, and drifting buoy and manned stations on Arctic sea ice. Efforts to extend and improve the data quality in COADS are continuing. Presently, COADS is the "best" available global set of marine surface data covering the 1854-1993 time period. The COADS is contained within the NCAR dataset ds540.1 (see Table 4.1). It includes 2 degree box statistical summaries of various observed variables and derived quantities on a year-month basis. These are presented in Table 4.2

As with land stations, the spatial distribution of ocean observations has changed significantly with time. Figure 4.2 depicts the number of 2 degree boxes with at least one observation for January from 1880 to 1979. This is summer in the southern hemisphere and represents the season with the best spatial distribution of observations. During the southern hemisphere winter (e.g., July), the number of 2 degree boxes with observations in high southern latitudes drops significantly. Indeed, if more than one observation (say, 5) is required to reduce sampling errors, the number of usable 2 degree boxes drops significantly in both the tropics and the southern latitudes (Fig. 4.3).

A dataset similar, but not identical, to COADS is the Meteorological Office Marine Data Bank (MDB) developed at the Hadley Centre, Meteorological Office, Bracknell, United Kingdom. COADS and MDB have many data in common but also each contains unique information. Long range plans are in place to blend COADS and MDB together. Climate research based on the MDB and COADS has contributed significantly to our understanding of climatic variations and of sampling biases that exist in these datasets that span many decades and numerous changes in sampling methods and technology. The quality of some climate change and climate variability estimates rely heavily on datasets like COADS and MDB. Improvements in these datasets are important for improved climate studies.

Satellite data are now being used to help define the surface boundary layer conditions. For example, AVHRR data are commonly used to augment in situ SST observations in order to provide more complete (both spatially and temporally) estimates. spatially and temporally) estimates. These data are combined using an optimal statistical method to produce gridded SST analyses. In a similar fashion, scatterometer data from ERS-1 and/or ERS-2 are used to improve estimates of wind over the oceans. In the near future information from the NASA scatterometer (NSCAT) will be used.

Subsurface Observations

In order to understand oceanic circulation, it is necessary to know the internal distribution of water mass (e.g., density) within the ocean. The distribution of ocean water density is defined by the water temperature, salinity, and pressure or depth. Measurements of these water properties are generally made with vertical profiling instruments lowered from research ships. These measure how assorted ocean properties vary as a function of depth. The most common is the temperature-depth profile. One of the first instruments developed specifically to measure this was the mechanical bathythermograph (MBT). This instrument is suspended from a cable and lowered from a stationary or slowly moving ship. Changes in water temperature are detected using xylene in a thin copper tube while pressure, which is closely related to depth, is sensed by a flexible diaphragm. The combined response of these sensors is recorded by a stylus scribing a mark on a gold plated glass slide. Following recovery of the MBT, the glass slide is removed and the scribed curve is read with a viewer containing depth and temperature grid scales. Sampling with this instrument is typically limited to the upper 100-200 meters depth. MBT measurements were prevalent for about three decades beginning in 1940. In the 1960's a new instrument, the expendable bathythermograph (XBT), was developed and superseded the MBT. Again the XBT was specifically designed to measure temperature-depth profiles, but it uses much different technology. Temperature is estimated by a temperature sensitive thermistor in the weighted nose of a small (~30 cm) torpedo shaped housing. The disposable XBT housing is dropped into the ocean. Freely uncoiling conductive wire maintains a connection with an on board recording device (a strip chart recorder or computer) so that temperature changes detected by the thermistor can be transmitted. The XBT descends through the sea water at an approximately constant rate, so depth is proportional to the time interval during the descent. After the full length of conductive wire is uncoiled, the XBT breaks free, ending the data collection. Nominally, most XBTs sample to a depth of about 400 meters but some XBTs are designed for deeper sampling (750 meters or more). When rapid sampling of large areas is required, specially designed XBTs may be deployed from low flying aircraft. XBTs are still commonly used because they are inexpensive and quite simple to use. Programs that arrange for deployment of XBTs from merchant vessels, while en route between ports, have been used successfully to collect much additional data at low cost.

The first systematic sampling (~1900) of the ocean's subsurface properties was done using self-closing sample bottles (e.g., Nansen and Niskin bottles) and calibrated thermometers on cables lowered from stationary ships. This sampling method, broadly called ocean station data, provides temperature and depth data at each location. The sample depths range from the surface to the maximum ocean depth. Not only is the depth of sampling greater than the BT technologies, but the recovery of water samples from discrete depths allows for laboratory determination of other ocean water properties. Primarily those properties have been salinity, nutrients (dissolved nitrates, nitrites, phosphates, and silicates), and dissolved oxygen. In this sampling scheme, ocean depth is determined by a thermometric technique whereby temperature differences between two thermometers (one vacuum sealed and protected from the ocean water pressure and the other exposed to the pressure) are used to estimate sample depth. Beginning in the mid-1960s, electronic instrumentation was introduced for use in conjunction with standard ocean station procedures. These instruments are called STDs or CTDs for the measurements of salinity or conductivity, temperature and depth (pressure). These instruments provide high resolution vertical profiles of temperature, salinity, and depth. In some cases, the high resolution data are augmented with water analyzed from sample bottles. Today the majority of ocean station data is obtained using CTD technology.

The temperature, salinity, and depth measurements from ocean stations are often used to calculate water density and, if data from several stations are available, a first order estimate of subsurface ocean current speeds can be made using the principles of geostrophic flow. The individual properties are also mapped and contoured in horizontal and vertical sections to describe water mass movements. The motivation for these types of studies is to develop an understanding of the general ocean circulation.

In recent years, water samples have been used to measure other ocean water properties. Dissolved gases, such as CO2 and N2O , are measured to improve our understanding of ocean-atmosphere exchanges of these compounds. Dissolved anthropogenic constituents such as helium-3, tritium, freon-11, and freon-12 are used as water parcel time stamps. Because constituents like these have known decay properties, the amount of time when a `water-parcel' was last in contact with the atmosphere may be approximated. These estimates provide time-history information on the slow-moving deep-ocean circulation and improve the understanding of water mass ventilation (i.e., water reaching the surface). There are numerous other ocean measurements that are made in support of biological and geological research that are not covered in this text.

The National Oceanographic Data Center (NODC) of NOAA is the primary source of ocean profile data. They have separate archives of MBT, XBT, ocean stations from bottle samples (SD), and both low and high vertical resolution CTD/STD stations. These historical archives typically have data within 2-3 years of the current date. The most recent data are often held by the groups who carry out the collection so that they can study the information prior to public release. Large field programs like the World Ocean Circulation Experiment (WOCE) and TOGA have archive centers responsible for collecting all data from many different ocean cruises. Following a reasonable amount of time, normally about two years, the data are submitted to NODC which provides data selection and data distribution services to all users. As part of ongoing data projects NCAR has the NODC archive (ds542.0). Smaller ocean station datasets include: Reid's selected deep ocean stations (ds543.0), Jenkin's tritium and helium-3 observations (ds544.0), Stalcup's observation from a warm core ring (ds545.0), Levitus' stations at standard levels (ds285.0), and Gordon's Southern Ocean stations (ds285.1). As a typical example, a vertical cross section of ocean temperatures from Europe to North America (Fig. 4.4) was prepared using Reid's dataset.

In ocean regions of particular interest (e.g., the Gulf Stream, the Kuroshio Current, the Antarctic Circumpolar Current) moored instrumented arrays are sometimes used to monitor oceanic flow and water properties. These arrays tend to be maintained for several years or less. The time series of observations have limited value for large spatial studies, but they do provide excellent data for energetic studies of ocean currents and calibration/comparison data for co-located ocean station survey data.

Unfortunately, placing instruments in the deep ocean (to depths of 5000 meters or more) is technically challenging and expensive. Therefore, these available records are few, but quite important to our understanding of ocean flow. One good source for ocean current data is the School of Oceanography at Oregon State University, Corvallis, Oregon. NODC also has current meter time series data holdings.


Analyses refer to gridded fields of properties derived from observed data, such as described in the previous sections. The majority of ocean analyses are for the sea surface. The reason for this is that most in situ ocean observations and satellite remote measurements apply only to the surface. Furthermore, understanding interactions between the ocean and atmosphere near the surface has long been recognized as important for better weather forecasting and has therefore received research emphasis. Analyses can be as simple as the interpolation of observed data onto a uniform grid. More complicated methods are used to develop grids during operational weather center forecasting procedures. Those procedures use real-time observed data in conjunction with previous output model grids and sophisticated data assimilation schemes (see Chapter 6).

Researchers involved in ocean-atmosphere comparative studies, and those who want an initial look at data often prefer analyzed grids because they are convenient to use. However, users should understand how the grids were produced and use them with appropriate caution. Some concerns might include: how noisy (spatial and temporal) were the original observations, what quality control and interpolation procedures were used. Ultimately, the question that remains for a user is "Have the procedures used to derive the gridded products adversely affected my scientific inferences?". Thorough metadata are critical to help scientists in answering this question.

Analyzed data span a full range of spatial and temporal scales. However, for most basic ocean research, gridded analyses that cover large portions of ocean basins with temporal resolutions ranging from several realizations per day to decades may be adequate. Analyses with higher temporal and/or finer spatial resolutions are often desired, but the lack of in situ observations, satellite remote sensing resolution scales, and operations weather center procedures preclude the development of these products.

Gridded climatologies, annual means, and monthly analyses provide first order estimates of ocean properties. The quality of these analyses generally varies with the density of the observational network used to derive them. Typically, the observational density, and thus the quality, is best over the northern hemisphere poleward of 20 degree N and is poorest over the southern oceans. Prior to the establishment of the TAO array, large areas of the tropical Pacific also had poor coverage. For most major oceanic variables (e.g., wind speed and direction, SST, atmospheric pressure, air temperature, humidity) these analyses are available on 2 degree x 2 degree (latitude x longitude) grids. Other variables (e.g., wave height and direction, cloud information, salinity, precipitation, sea ice limits, etc.) are more poorly represented. Derived quantities like heat fluxes, radiation fluxes, and momentum fluxes are also not as good as desired. Nevertheless, in certain regions where adequate sampling is available, reasonable estimates are available and have proved useful in climate research. Examples of datasets archived at NCAR that fall into this category include: Trenberth's global wind stress climatology based on ECMWF analyses (ds110.1; see Chapter 6), Hellerman's monthly global wind stress (ds232.0), Shea's global monthly SST climatology (ds289.0), Legates' global air temperature and precipitation monthly climatology (ds236.0), Esbensen's global wind stress and heat budget climatology (ds209.0), Petty's global ocean precipitation characteristics (ds541.2), Oberhuber's global climatological atlas (ds541.0) based on COADS (ds540.0/ds540.1), and the Atlas of Surface Marine Data (ds541.1; da Silva et. al. 1994) also derived from COADS.

Sub-surface analyses are very important for oceanographic research, especially for large scale comparative studies and model initialization and verification. Several sub-basin scale analyses have been made (NCAR datasets ds286.0, ds278.0, ds544.0, and ds285.1). Two global datasets have been developed by Sydney Levitus at NODC. The first, which was very popular, was published in 1982. That dataset is now superceded by the World Ocean Atlas 1994. Data used to contruct the 1994 Atlas are the largest set of ocean profile data ever assembled. This multi-level atlas provides annual, seasonal, and monthly analyses on a 1 degree grid. Furthermore, the profile data and 5 degree gridded analyses are also available. The gridded variables include temperature, salinity, dissolved oxygen, apparent oxygen utilization, oxygen saturation, and nutrients (NO3, SIO2, PHO4). An annual temperature anomaly analysis has also been derived from these observations. The analyses are limited to the 0-500m depth range and cover the 1960-1990 time period. This dataset is available from both NODC (see Appendix A) and from NCAR (ds285.0). Table 4.3 provides a brief overview of this dataset.

In support of tropical Pacific Ocean studies and associated coupled ocean-atmosphere studies, NMC now makes available the first operational sub-surface ocean analyses. Using observational data from ships and the TOGA/TAO moored buoys, an ocean model produces analyzed grids of sea surface temperature and ocean current velocity on a weekly basis for the tropical Pacific Ocean (ds277.1). Analyses like these are presently only possible in regions with adequate subsurface in situ measurements. These analyses represent a new frontier for ocean science and should improve with time and experience.

When the research focus changes from mean conditions to long term trends or interannual variability, datasets spanning long time periods are required. Typically, gridded monthly time series, often called year-month time series (12 monthly grids per year), are needed for this type of research. The COADS provides a long year-month history by statistically summarizing in situ data in 2 degree x 2 degree areas, for all available data, for the period 1854-1992. The variables summarized are those commonly taken at synoptic hours on board ships (e.g., wind speed and direction, SST, air temperature, atmospheric pressure, relative humidity, and cloud type). In general, these summaries are of higher quality and provide better global coverage in the more recent years. SST is the most frequently analyzed variable. Important improvements have been made in these analyses beginning in 1970 by incorporating sea-ice edge location and using optimal interpolation methods to estimate various quantities in data sparce regions. In these recent analyses, the SST interpolation is constrained to correctly match the freezing temperature of sea water at the sea ice edge location. This greatly improves the representation of SST in high latitude regions which often lack sufficient in situ observations. Another advancement in SST analyses became possible in 1980 with the availability of satellite AVHRR data. To date, our best available SST analyses are derived using in situ observations as ground truth, in combination with AVHRR data to provide patterns to fill data sparse regions, and realistic sea ice edge specification (also obtained from satellite measurements) to constrain the analyses at high latitudes.

Confidence in these techniques has advanced to the point that weekly 1 degree x 1 degree global SST grids are now available in near real time. These analyses improve our ability to monitor changes in SST (e.g., the El Nino and La Nina phenomena). Improved SST analyses are also available for the 1950-1981 period. Using empirical orthogonal functions (EOFs) determined from the 1982-1992 period as basis functions the in situ observations for earlier time periods have been interpolated. The horizontal structure and length scales within these grids are now more consistent with analyses from later time periods. These various gridded data are contained within NCAR ds277.0.

Many other analyses that fully or partially rely on satellites are available. Several have been mentioned here, but the interested reader is referred to the monthly publication, Climate Diagnostic Bulletin, U.S Dept. of Commerce, NOAA, National Weather Service, National Meteorological Center . In that publication, analyses are graphically displayed and brief discussions are often presented. Monthly sea level displacement from the mean is derived using the European Space Agency ERS-1 satellite altimeter and tide gauge data from islands in the Pacific Ocean. Global ocean dynamic topography, which reflects relative ocean circulation, is derived on a two week or monthly basis using the TOPEX/POSEIDON satellite altimeter data with a set of ocean tidal and geoid models. Other promising developments include tropical rainfall analyses based on outgoing long-wave radiation measurements and surface wind speed and directions as determined by scatterometer measurements taken from satellites. Developing ocean analyses such as these have potentially important benefits.

Analyses, done as part of operational weather center forecasting procedures, have proven to be valuable research datasets (see Chapter 6). These analyses offer relatively high temporal and spatial resolution. Major centers (e.g., NMC and ECMWF) now provide analyses at intervals as often as four times per day at many atmospheric levels including the surface boundary layer for many variables including SST, air temperature, pressure, winds, humidity, fluxes, and precipitation. A wide variety of real time ocean and atmospheric observations are used in each analysis. The data are subjected to consistency checks with previous analyses. Grids of this type became available from NMC in 1976, but it was not until 1979, during FGGE (see Chapter 8), that global analyses first became viable. Even though 18 years of data are available it is not suitable for long term trend evaluation. During this time period there have been significant developments (improvements) in numerical techniques and computational coding of model dynamics. These developments have led to model changes that cause changes in the analyzed grid products not related to climate changes. The problem is how to separate changes resulting from changing procedures from real climate change. To address this problem, four "reanalysis" projects are underway. Chapter 6 presents more details on this project.

Sea-Ice, Sea Level, Topography Data

The previous sections identify some important types of ocean data, however there are often other data types required for oceanographic research. The global ocean bottom topography (A complete listing of topography datasets at NCAR appears in Chapter 9.) and geographical limits of the ocean basin are defined with ocean bathymetry datasets. A 1 degree resolution set from ds750.1) gives a good representation at low resolution while the ETOPO5 (Earth Topography 5 minute) dataset from the National Geophysical Data Center (ds759.1) presents ocean depth and land elevation at a 5'x5' resolution. Other higher resolution datasets, for limited regions are available from the Defense Mapping Agency, Department of Defense. The mean position and the variability of sea ice limits are also important for some oceanographic studies. To fill this requirement NCAR provides some basic datasets identified as ds233.0, ds234.0 and ds270.2 in Table 4.1. These datasets are derived from higher resolution archives at the National Ice Center (formerly the U.S Navy Joint Ice Center) and the NSIDC. Under certain circumstances (e.g., coastal salinity studies) the freshwater flux into the coastal ocean is required. NCAR has a small archive of selected world river flow rates (ds552.0). Other river flow data are available from U.S Geological Survey (see Appendix A). A small selection of historical sea level time series data is available at NCAR. The records from 1800-1987 are available in ds252.0. More modern and comprehensive records can be obtained from the WOCE Sea Level Data Center at the University of Hawaii (see Appendix A).

Table 4.1
Selected NCAR Oceanographic Datasets
ds110.1Trenberth, Global Wind Stress Climatology Based on ECMWF1741980-89
ds209.0Esbensen, Global wind stress and heat budget climatology 9 
ds209.1Weare, Tropical Pacific Ocean yr-mo heat budget23 1957-79
ds209.2semer and Hasse, Bunker Climate Atlas of the North Atlantic25 
ds209.3Hastenrath, Heat budget atlas for the tropical Atlantic 5 
ds230.0OBrien, FSU trop. Pacific Indian Ocean. yr-mon wind str. 201961-92
ds231.0Wyrtki, Tropical Pacific yr-mon surface wind stress9 1947-73
ds232.0Hellerman, GFDL Monthly Global wind stress 3  
ds232.1Harrison, Climatological mean wind stress, global ocean 19  
ds236.0Legates, Global air temp. and precip., monthly climatology 84  
ds250.0Pacific hourly, daily, and monthly sea level heights 51 1901-87
ds251.0EPOCS, Equatorial Pacific Ocean Climate Studies datasets 94 1950-79
ds252.0PSMSL Permanent Service Mean Sea Level station data 241800-1987
ds253.0Air and Surface water co2 and n2, obs., global ocean21977-90
ds254.0Najjar, Global ocean nutrient grids (po4, no3, sio2), 1x145  
ds255.0 Hansen, AOML EPOCS drifting buoy position and SST 9 1979-84
ds256.0MEDS, Global surface drifting buoy dataset15001978-93
ds256.1PMEL/TOGA-TOA Atlas moored buoys, EPOCS moored buoys421980-91
ds256.2Colony, Arctic SLP and T. from Ice Buoys, 2 x daily251 1979-90
ds257.0MEDS, Canadian West and East Coast sst and salinity101914-85
ds258.0Scripps pier west coast temp. and salinity31916-90
ds258.0FOSDIC marine sfc/upper air 31916-90
ds259.0 utrients, Chlorophyll, Primary Production, etc 4 1959-99
ds271.0 Oort and Yi, Global yr-mon SST grids from COADS 3931880-79
ds272.0Sadler, Tropical Marine Climate Atlas, sst, slp, wind stress 10  
ds274.0Rasmussen, Tropical Pacific yr-mo sst, wind, t-air 2041946-76
ds277.0NMC yr-mo, weekly, and climate SST (blend+OI+EOF) 401950-94
ds277.1Leetmaa, NMC, wkly, trop. Pacific, subsurf. U, V, and T 520 1991-94
ds278.0Bauer, Global ocean near surface monthly and annual climatology70  
ds279.0Samuels Cox, GFDL Dataset Atlas for Oceanographic Model. 165  
ds280.0 Meehl, Global long-term mean ocean surface currents .2  
ds285.0Levitus (NODC), World Ocean Atlas: 1 degree grids and obs.32001900-93
ds285.1Southern Ocean Atlas, gridded (1x2) and observed stations 25 1900-75
ds286.0 Fukumori, The Hydrography of the North Atlantic Ocean... 18  
ds287.0 GFDL MOM Model Climatological Datasets 310  
ds289.0 Shea et. al (1992), Global Monthly SST Climatology (2 degree; via ftp)1  
ds289.1 Bottomley's Global Ocean Surface Temperature Atlas 68  
ds474.0 Colony, Russian Ice Stations Meteorological Obs., 6-hrly 10 1950-90
ds533.0 USSR Marine Ship Archive, surface marine obs. 2510 1888-1991
ds541.2 Petty's global ocean precipitation characteristics  
ds535.0 Observations from Ocean Weather Ships 470 1941-91
ds541.0 Oberhuber, An atlas based on COADS, global, climatology 75  
ds541.1 da Silva, Atlas of Marine Surf. Data, clim. and anomalies 3000 1945
ds542.0 NODC XBT, MBT, SBT, SD, and STD/CTD station data archive 2763 1900-92
ds543.0 J.L. Reid, Selected deep ocean stations (7000 obs) 18 1900-87
ds544.0 Jenkins, Tritium and helium-3 from TTO-NAS and NATS .3 1981
ds545.0 Hydrographic Data from Warm Core Ring 82-B, Stalcup et al. 9 1982
ds545.1 Gulf Stream Anatomy Hydrographic Survey - Fall 1988, Rossby 4 1988
ds552.0 UNESCO selected river flow rates 2 1800-1972
ds726.0 Atlas, SEASAT scatterometer derived wind stress, dealiased 200 1978
ds726.1 Chelton, SEASAT scatterometer derived wind stress 1 1978
ds727.1 Fu and Zlotnicki, JPL GEOSAT Gridded Data Track File 520 1986-88
ds744.0 ESA ERS-1 satellite scatterometer data. 8500 1991-1995
 Ocean Depth and Land Elevation
ds750.1 Scripps/RAND Corp. global land elevation / ocean depth 1  
ds759.1 ETOPO5 5 minute gridded world elevation / ocean depth 57  
ds754.0 Global 10 minute Elevation Dataset from the Navy 56  
 Comprehensive Ocean Atmosphere Dataset (COADS)
*global ocean surface data from ships and buoys*
ds540.0 CMR, Compressed Marine Reports, 72 million obs. 1723 1854-1979
ds540.0 LMRF, Long Marine Reports, 45 million obs. 2700 1980-93
ds540.1 MSTG, Monthly Summary Statistics, in 2x2 squares 1600 1854-1993
 Datasets from Operational Atmospheric Analyses
ds082.1 NMC 2.5x2.5 Global Grids, surface subset, 2xdaily 3000 1976-95
  NMC 2.5x2.5 Global Grids, yr-mon grids 100 1976-95
ds084.5 NMC Med. Range Forecast Model Flux Archive, 4xdaily 2200 1990-95
ds240.0 U.S. Navy Fleet Numerical Oceanography Center 8123 1961-91
ds110.0 ECMWF 2.5x2.5 Global Grids, near surface subset 663 1980-89
ds111.1 ECMWF/TOGA Adv. Oper. Dataset, 4xdaily, 1x1 resol. 29000 1985-95
ds111.2 ECMWF/TOGA Basic 2.5x2.5 Analysis, 2xdaily 15000 1985-95
ds108.0 Australian National Meteorological Research Centre 1720 1972-89
 Sea Ice
ds233.0 Walsh, Arctic Monthly Sea Ice Concentration 5 1953-88
ds234.0 Ropelewski, Antarctic Monthly Ice Area 1 1973-90
ds270.2 Alexander and Mobley, Monthly Average SST+Ice-Pack Limits 2  

Table 4.2
COADS: Variables and Derived Quantities
Slutz et al, 1985
# betaObserved Variable
1S sea surface temperature
2A air temperature
3W scalar wind
4U vector wind eastward component
5V vector wind northward component
6P sea level pressure
7C total cloudiness
8Q specific humidity


9R relative humidity
10D S - A = sea-air temperature difference
11E (S - A)W = sea-air temperature difference wind magnitude
12F Qs - Q (saturation Q atS) -Q
13G FW - (Qs)W (evaporation parameter)
14X WU
15Y WV (14-15 are wind stress parameters)
16I UA
17J VA
18K UQ
19L VQ (16-19 are sensible and latent heat transport parameters)

1d mean day-of-month of observations
2h hour statistic of observations
3x mean longitude of observations
4y mean latitude of observations
5n number of observations
6m mean
7s standard deviation
80 0/6 sextile (the minimum)
91 1/6 sextile (a robust estimate of m-1s)
102 2/6 sextile
113 3/6 sextile (the median)
124 4/6 sextile
135 5/6 sextile (a robust estimate of m+1s)
146 6/6 sextile (the maximum)

Table 4.3
Ocean Climate Laboratory (NODC)
Observational dataSize (MB)
1All the observed level profile data2278
2 All profile data interpolated to 33 standard levels 962

WOA 1 degree analyzed data
Parameter AnnualSeason Month Anomaly
Temperature X X X X
Salinity X X X 
Dissolved oxygen X X    
Apparent Oxygen Utilization X X  
Oxygen saturation X X X 
Phosphate X   
Nitrate X   
Silicate X   
Annual - composite of all data regardless of season or year; 33 levels: 17MB/file
Season - data composite based on conventional NH seasons; 33 levels: 17MB/file
Month - data composite for each month. 19 standard levels (0-1000 m): 10MB/file
Anomaly - 1960-90; 14 standard lvls.; (0-500m): 7MB/file

WOA 5 degree analyzed data
Parameter Annual Season
Temperature X X
Dissolved oxygen X X
Apparent Oxygen Utilization X X
Oxygen saturation X X
Nitrate X  
Phosphate X 
Silicate X 
Means, standard deviation, and number of aobservtions are available in separate files: 33 standard levels; 0.7MB/file

Ocean-Atmosphere Boundary Layer Data
Subsurface Observations
Sea-Ice, Sea Level, Topography Data

An Introduction to Atmospheric and Oceanographic Datasets
[Next] [Prev] [Contents] [Top]