Ancillary Data Sources

1. Introduction

The algorithms for retrieval of ocean color radiometry and sea surface temperature from satellite radiometry make use of a number of ancillary data sources, in addition to the sensor observed radiometry. Examples include meteorological data (windspeed, surface pressure, relative humidity) and concentrations of atmospheric gases (water vapor, ozone, nitrogen dioxide). In general, the optimal ancillary data is not available when the satellite radiometry is first acquired, so the NASA OBPG processing uses best available ancillary data sources for processing satellite data in near realtime, and then reprocesses the data about one week later to refine the products using the optimal ancillary data. The primary sources of ancillary data used in NASA ocean color and SST processing are described here.

Section 2 will quickly cover the ancillary data sources used to make the ocean color and the sea surface temperature products. Section 3 will cover the ancillary products in greater detail.

2. Ancillary Data for Ocean Color and SST Products

Ancillary data for Ocean Color Products

The major sources of ancillary data required to process ocean color data are meteorological, ozone, nitrogen dioxide (NO2), and sea ice.

  • Meteorological data (l2gen: MET1, MET2, MET3). Several pieces of meteorological data are used to properly remove atmospheric and surface effects from the radiances to compute the water-leaving radiances in the level-1 to level-2 processing program (l2gen). Surface wind information is used to estimate and remove the amount of sun glint, surface pressure is used to determine the amount of Rayleigh radiance, surface relative humidity is used to help in choosing the best aerosol model, and the precipitable water is used to determine the atmospheric transmittance. Four sources of meteorological data are used to create the best near-real-time and refined ocean color products: Climatology, Forecast meteorology data, Real-time data, and Reanalysis 2 improved data. The climatology data is always available but has the lowest quality for use on a specific day. The Reanalysis 2 improved meteorological data is made from the most observations, using a consistent model and has the highest quality, but it has a latency of 2 - 6 weeks.

    ECMWF meteorological data alternative - As an alternative, the l2gen code has the option of accepting and using ECMWF Interim Analysis files of meteorological and ozone data. Due to the ECMWF data use policy, users will have to acquire the datasets directly from ECMWF (see the ECMWF section for information on doing this).

  • Ozone data (l2gen: OZONE1, OZONE2, OZONE3). Atmospheric ozone also plays a significant role in the derivation of the correct water-leaving radiance. The concentration of ozone affects the visible light absorption, most strongly in the green. We have an ozone climatology file and real-time ozone files for processing use. As a backup source, the TOAST ozone data is also available.
  • Nitrogen Dioxide (NO2) data (l2gen: NO2FILE). Like ozone, but to a lesser extent, nitrogen dioxide (NO2) accounts for some atmospheric absorption of the visible radiation, mainly in the blue (Robinson et al. 2007). Currently, we only have a NO2 climatology for processing. The daily NO2 from different satellite missions (GOME, SCIAMACHY, and OMI) is being evaluated to make a daily NO2 record and to possibly use it in the operational processing.
  • Sea Ice Data (l2gen: ICEFILE). The sea ice concentration is used to locate water regions that are ice covered and should be avoided when computing the Photosynthetically Active Radiation (PAR), along with land and high glint regions. A climatology and daily ice concentration data are available.

When the OBPG gets a satellite dataset, the best available ancillary dataset among the MET and ozone data (climatology or real-time) are used to make an ocean color product for timely dissemination - the quicklook product. When the best possible ancillary data are available, or after 45 days have passed, the ocean color data is processed again with the best available ancillary data to make the final product - the refined product. The data selection sequence will use the topmost data source in the list below:

  • Reanalysis-2 improved meteorological data
  • Real-time meteorological data surrounding the data time and within 6 hours of the time
  • Forecast NCEP meteorological data
  • Real-time meteorological data in neighborhood up to 12 hours of the time
  • Climatology

How to specify ancillary meteorological and ozone data in l2gen. The l2gen processing program has 3 inputs to accommodate the meteorological data: MET1, MET2, and MET3, and the ozone data: OZONE1, OZONE2, and OZONE3. This is provided so that the ancillary file parameters may be interpolated in time to the time the satellite data was acquired. The 3 files are selected so that the times of the ancillary data encompass the times in the satellite granule.

  1. Climatology use. If climatology is selected, the file name for it need only be specified for MET1 or OZONE1.
  2. Use one ancillary time for entire granule. This can be done, if desired by specifying the same ancillary file name for MET1, 2, and 3 or OZONE1, 2, 3.
  3. Interpolate 2 ancillary times. If both the start and end time of the satellite granule fall between 2 ancillary data times, specify MET1 or OZONE1 as the first ancillary time and MET2 or OZONE2 as the second ancillary time. Repeat the second ancillary time in MET3 or OZONE3.
  4. Interpolate 3 ancillary times. If one of the ancillary times falls in-between the start and end times of the satellite granule, specify MET1 or OZONE1 as the ancillary time before the granule start time, MET2 or OZONE2 as the ancillary time between the granule start and end time, and MET3 or OZONE3 as the ancillary time after the granule end time.

OLCI tie point meteorology and ozone data. In the special case of OLCI instrument data from the ESA Sentinel 3 spacecraft, every level-1 granule has a set of meteorological and ozone parameters that are specifically created to cover the region seen by the granule. This tie point meteorological file can be used in the processing or the other standard meteorological and ozone files can be specified. To specify the tie point file, use the name as the MET1 and/or OZONE1 inputs to l2gen. Alternatively, the keyword 'OLCI_TIE_METEO' can be specified for MET1 and/or OZONE1 and l2gen will take the correct tie point met file name from the manifest file for that granule (note that the manifest file is specified as IFILE for OLCI processing by l2gen). The tie point data is supplied by ECMWF.

Ancillary Data for the Sea Surface Temperature (SST) Products (l2gen: SSTFILE)

The SST is derived using the infrared window bands that are on some of the ocean color instruments and on the AVHRR instrument on the TIROS and NOAA meteorological satellites.The 11 and 4 micron window bands can be used to derive the SST. A SST reference file is used as an aid in estimating the SST from the infrared band brightness temperature values. The reference file can be a climatology, a daily, preliminary SST estimate or a final estimate. Each data source, like the meteorological data mentioned above, does a trade-off between timeliness and quality.

3. Detail of the ancillary sources

Meteorological Climatology

The climatology of the meteorological data currently is made from the monthly weighted averages of the reanalysis 2 improved meteorological data. Since the ocean color instruments generally take data at local noon, the weighting of the individual 6-hourly data times is done so that the local noon data has the greatest weight.

Forecast Meteorological data

The forecast meteorological data is taken from the NCEP Global Forecast System (GFS) data at forecast times (12 and 15 hour forecasts) that will allow it to be available in time to perform quick-look processing. This data source should normally be available for all ocean color processing. However, if it is not, climatological data will be used.
Parameters: mean sea level surface pressure, 10 meter above ground level winds (U and V), 1000 mb relative humidity, and precipitable water vapor.
Source organization: National Center for Environmental Prediction (NCEP).
Spatial resolution: 0.25 x 0.25 degrees (1440 x 720 grid)
Temporal resolution: Every 3 hours starting 0000 GMT.
Latency: about 7 hours.
Time covered: 1978 - present.
Reference: Derber, J.C., D.F. Parrish, and S.J. Lord, 1991:L The new global operational analysis system at the National Meteorological Center. Wea. Froecasting, 6, 538-547).
Naming convention: form: NYYYYDDDHH_MET_NCEP_1440x720_fhh.hdf, with YYYY=year, DDD=day-of-year, HH=hour-of-day, 00 to 21 GMT by 3, hh=forecast amount in hours, either 12 or 15 hour forecasts. example: N201603003_MET_NCEP_1440x720_f15.hdf

Real-time Meteorological data

The real-time meteorological data is derived from fields in the Global Data Assimilation System (GDAS) datasets available from the National Center for Environmental Prediction (NCEP). These analyses are available about 7 hours after the observation time.
Parameters: mean sea level surface pressure, 10 meter above ground level winds (U and V), 1000 mb relative humidity, and precipitable water vapor.
Source organization: National Center for Environmental Prediction (NCEP).
Spatial resolution: 1 x 1 degrees (360 x 181 grid)
Temporal resolution: Every 6 hours starting 0000 GMT.
Latency: about 7 hours.
Time covered: 1978 - present.
Reference: Derber, J.C., D.F. Parrish, and S.J. Lord, 1991:L The new global operational analysis system at the National Meteorological Center. Wea. Froecasting, 6, 538-547).
Naming convention: form: NYYYYDDDHH_MET_NCEP_6h.hdf with YYYY=year, DDD=day-of-year, HH=hour-of-day, 00 to 18 GMT by 6. example: N201603003_MET_NCEP_6h.hdf

Reanalysis 2 improved meteorological data

The NCEP Reanalysis 2 product has more consistency over time (one fixed model is used to analyze the data) and includes more observations in it's analysis (it waits for more observations to arrive) than the real-time data does, at the cost of timeliness (data is available with a 3 - 4 day delay). This data is used to improve the quality of the real-time data (real-time has 1 degree resolution and the Reanalysis 2 has 2.5 degree, so the real-time fields are adjusted to match the reanalysis).
Parameters: mean sea level surface pressure, 10 meter above ground level winds (U and V), 1000 mb relative humidity, and precipitable water vapor.
Source organization: National Center for Environmental Prediction (NCEP).
Spatial resolution: 1 x 1 degrees (360 x 181 grid)
Temporal resolution: Every 6 hours starting 0000 GMT.
Latency: 3 - 4 days.
Time covered: 1996 - present.
Reference: Kanamitsu, M., W. Ebisuzaki, J Woollen, S.-K. Yang, J. J. Hnilo, M. Fiorino, and G. L. Potter, 2002: NCEP-DOE AMIP-II Reanalysis (R-2), Bull. Amer. Meteor. Soc., 83, 1631-1643.
Naming convention: form: NYYYYDDDHH_MET_NCEPR2_6h.hdf with YYYY=year, DDD=day-of-year, HH=hour-of-day, 00 to 18 GMT by 6. example: N201603003_MET_NCEPR2_6h.hdf

ECMWF Interim Analysis data

The ECMWF-supplied meteorological and ozone data are a distinct source of ancillary data that can be used in ocean color processing. The OBPG version of this data is a subset acquired from ECMWF without local modification. Note that as the relative humidity at the surface is not provided, it is derived from the 2 m temperature, dewpoint temperature, and the surface pressure.
Parameters: mean sea level pressure, 10 meter above ground level winds (U and V), total ozone, 2 m dewpoint temperature, 2 m temperature, surface pressure, and precipitable water vapor.
Source organization: European Centre for Medium-Range Weather Forecasts (ECMWF).
Spatial resolution: 0.75 x 0.75 degrees (480 x 241 grid)
Temporal resolution: Every 6 hours starting 0000 GMT.
Latency: 1 to 2 months.
Time covered: 1979 - present.
Reference: Dee, D. P., Uppala, S. M., Simmons, A. J., Berrisford, P., Poli, P., Kobayashi, S., Andrae, U., Balmaseda, M. A., Balsamo, G., Bauer, P., Bechtold, P., Beljaars, A. C. M., van de Berg, L., Bidlot, J., Bormann, N., Delsol, C., Dragani, R., Fuentes, M., Geer, A. J., Haimberger, L., Healy, S. B., Hersbach, H., Hólm, E. V., Isaksen, L., Kållberg, Köhler, M., Matricardi, M., McNally, A. P., Monge-Sanz, B. M Morcrette, J.-J., Park, B.-K., Peubey, C., de Rosnay, P., Tavolato, C., Thépaut, J.-N. and Vitart, F. (2011), The ERA-Interim reanalysis configuration and performance of the data assimilation system. Q.J.R. Meteorol. Soc., 137: 553-597. doi: 10.1002/qj.828
Naming convention: (these files are user-acquired so no naming is done).

Procedure for acquiring the ECMWF Interim Analysis data for use as an ancillary meteorological source:

  1. Go to the ECMWF data server on your browser:
    http://apps.ecmwf.int/datasets/data/interim_full_daily/levtype=sfc/
    (you will need to set up an account to get data)
  2. Select time range of data and times in a day desired, and select step 0 to get analysis data, not forecast. Also, select at least the following 8 parameters:
    • 10 metre U wind component
    • 10 metre V wind component
    • 2 m dewpoint temperature
    • 2 m temperature
    • surface pressure
    • Mean sea level pressure
    • Total column ozone
    • Total column water vapor
    With all this selected, press the 'Retrieve NetCDF' and in the next page, 'Retrieve Now'
  3. Separate into single-time files.
    If you just acquired 1 time above, the file is ready for use in l2gen. However, if you acquired multiple times, you need to make single time files for use in l2gen. This can be done using netcdf program ncks:
    ncks -d time, <time index> <start file> <end file>
    For example to get the 1st time in the file, do:
    ncks -d time,0 file_in.nc file_out.nc
    file_out.nc is now ready for use in l2gen

Ozone Climatology

We have a daily climatology of ozone based on the Aura OMI real time Aura OMI ozone data from 2004 - 2013.

Real Time Ozone

The real time ozone data is derived from TOMS-like instrument data. Currently, we are using data from the AURA OMI instrument (AURAOMI), and prior to 2005, we used data from the Nimbus-7 (N7TOMS) and Earth Probe (EPTOMS) TOMS instruments.

We apply a normalization to all the ozone data based on a de-trended merged SBUV record
Parameters: total column ozone, Dobson Units
Source organization: NASA TOMS project.
Spatial resolution: 1.25 x 1. degrees (288 x 180 grid)
Temporal resolution: Daily.
Latency: 1 to 2 days.
Time covered: (year/day-of-year used by the processing system) N7TOMS: 1987/305 - 1986/174, EPTOMS: 1996/305 - 2005/348, AURAOMI: 2005/349 - present.
Reference: R. McPeters, R. Stolarski, and S,. Frith, A Long-Trem Merged Global Ozone Data Set, ISPRS Proceedings, 2011, (http://www.isprs.org/proceedings/2011/isrse-34/211104015Final00262.pdf).
Naming convention: form: NYYYYDDD00_O3_III_24h.hdf with YYYY=year, DDD=day-of-year, III= instrument, N7TOMS, EPTOMS or AURAOMI. example: N201603000_O3_AURAOMI_24h.hdf.

Real Time TOAST Ozone

As a backup data source, the Total Ozone Analysis using SBUV/2 and TOVS (TOAST) data is available. This should only be used should the standard TOMS ozone not be available.
Parameters: total column ozone, Dobson Units
Source organization: NOAA Office of Satelite and Product Operations.
Spatial resolution: 1.25 x 1. degrees (288 x 180 grid)
Temporal resolution: Daily.
Latency: 1 to 2 days.
Time covered: 2004 - present.
Reference:http://www.ospo.noaa.gov/Products/atmosphere/toast/index.html
Naming convention: form: SYYYYDDD00DDD23_TOAST.OZONE with YYYY=year, DDD=day-of-year, the 00 and 23 are the range in hours covered by the data. example: N20160300003023_TOAST.OZONE.

NO2 climatology

We currently use a monthly climatology of NO2 based on the Aura OMI data.  Currently, only the climatology is used in ocean color processing. The climatology is created from the daily OMI NO2 files.  The daily files of NO2, as it comes from the DAAC has many gaps and only has good coverage when averaged into monthly files.
Parameters: total NO2, tropospheric NO2, stratospheric NO2.
Source organization: Goddard Earth Science Data Information and Services Center (GES DISC).
Spatial resolution: 0.25 x 0.25 degrees (1440 x 720 grid)
Temporal resolution: Monthly, not year-specific.
Latency: None.
Time covered: Any time. made from data in the years 2004 - 2013.
Reference: P. Stammnes, (Ed), OMI Algorithm Theoretical Basis Document, Volume III, Clouds, Aerosols, and Surface UV Irradiance, http://eospso.gsfc.nasa.gov/sites/default/files/atbd/ATBD-OMI-03.pdf

Ice Concentration Climatology

The ice concentration climatology is a monthly average ice concentration for all months in the years 1997 to 2007, accumulated from the NSIDC monthly ice concentration data. For times outside these bounds, the closest time is used.

Parameters: ice concentration from 0 - 255 (0 - 100% coverage) Source organization:National Snow and Ice Data Center (NSIDC) http://nsidc.org Spatial resolution: North and South polar projections, variable resolution. 488x304 North, 332x316 South Temporal resolution: Grids for each month of the year
Latency: None.
Time covered: All time - a climatology
Reference: Comiso, J. 1990, updated 2005. DMSP SSM/I daily and monthly polar gridded sea ice concentrations. Edited by J. Maslanik and J. Stroeve. Boulder, Colorado USA: National Snow and Ice Data Center. Digital media.

Daily Ice Concentration

The daily ice concentration is also from the NSIDC.

Parameters: ice concentration from 0 - 255 (0 - 100% coverage) Source organization:National Snow and Ice Data Center (NSIDC) http://nsidc.org Spatial resolution: North and South polar projections, variable resolution. 488x304 North, 332x316 South Temporal resolution: Grids for each month of the year
Latency: None.
Time covered: 26 Oct, 1978 - present
Reference: Meier, W., F. Fetterer, K. Knowles, M. Savoie, M. J. Brodzik. 2006, updated quarterly. Sea ice concentrations from Nimbus-7 SMMR and DMSP SSM/I passive microwave data. Boulder, Colorado USA: National Snow and Ice Data Center. Digital media.
Naming convention: form: NYYYYDDDHH_SEAICE_NSIDC_24h.hdf with YYYY=year, DDD=day-of-year, HH=hour. example: N201603000_SEAICE_NSIDC_24h.hdf.

SST Climatology Data

The SST climatology is a monthly climatology of the SST over the world.
Parameters: Sea Surface Temperature (degrees C) average for each month of the year
Source organization:JPL, NSIPP AVHRR Pathfinder, http://podaac-www.jpl.nasa.gov/ or http://podaac.jpl.nasa.gov/products/product112.html
Spatial resolution: 4096 x 2049 lat, lon grid, about 9 km grid size (at equator)
Temporal resolution: Grids for each month of the year
Latency: None.
Time covered: All time - a climatology
Reference: http://podaac.jpl.nasa.gov/products/product112.html

SST Preliminary Daily Analysis data

The SST preliminary analysis is a daily file of Sea Surface Temperatures. It uses 1 day of in situ (ship and buoy) observations and AVHRR satellite data in the optimum interpolation (OI), with a satellite bias correction using 7 days (one sided) of data and without smoothing of the EOT modes. This reference SST, the instruments' infrared-window brightness temperatures, and coefficient files (latitude zone and month deliniated) are used to determine the SST.
Parameters: Sea Surface Temperature (°C), SST anomaly (°C), SST Standard deviation estimated error (°C), and Sea ice concentration (%)
Source organization:NOAA/National Climatic Data Center
Spatial resolution: 720 x 1440 (0.25°)
Temporal resolution: Daily
Latency: 1 day
Time covered: 1981 - current
Reference: Reynolds, R.W., T.M. Smith, C. Liu, D.B. Chelton, K.S. Casey, and M.G. Schlax, 2007: Daily High-Resolution-Blended Analyses for Sea Surface Temperature. J. Climate, 20, 5473-5496, doi: 10.1175/2007JCLI1824.1
Naming convention: form: NYYYYDDD_SST_OIV2AV_24h.nc with YYYY=year, DDD=day-of-year; example: N2016030_SST_OIV2AV_24h.nc

SST Final Analysis data

The SST Final Analysis uses 3 days (centered) of in situ and satellite data in the OI, with a satellite bias correction using 15 days (centered) of data and smoothing of the EOT modes over 5 days (centered), to make a better estimate of the SST. The parameters are the same as the preliminary SST files (see above) and replace the preliminary files on our archive with the same name after a 16-day delay.
Latency: 16 days

References

Derber, J.C., D.F. Parrish, and S.J. Lord, 1991:L The new global operational analysis system at the National Meteorological Center. Wea. Froecasting, 6, 538-547.

Kanamitsu, M., W. Ebisuzaki, J Woollen, S.-K. Yang, J. J. Hnilo, M. Fiorino, and G. L. Potter, 2002: NCEP-DOE AMIP-II Reanalysis (R-2), Bull. Amer. Meteor. Soc., 83, 1631-1643.

Dee, D. P., Uppala, S. M., Simmons, A. J., Berrisford, P., Poli, P., Kobayashi, S., Andrae, U., Balmaseda, M. A., Balsamo, G., Bauer, P., Bechtold, P., Beljaars, A. C. M., van de Berg, L., Bidlot, J., Bormann, N., Delsol, C., Dragani, R., Fuentes, M., Geer, A. J., Haimberger, L., Healy, S. B., Hersbach, H., Hólm, E. V., Isaksen, L., Kållberg, P. Köhler, M., Matricardi, M., McNally, A. P., Monge-Sanz, B. M. Morcrette, J.-J., Park, B.-K., Peubey, C., de Rosnay, P., Tavolato, C., Thépaut, J.-N. and Vitart, F. (2011), The ERA-Interim reanalysis configuration and performance of the data assimilation system. Q.J.R. Meteorol. Soc., 137: 553-597. doi: 10.1002/qj.828