NASA Logo
Product Format and Content Changes for R2009

Product Format and Content Changes for R2009

Level-1 Format Changes

The lower-level formats for all sensors remain effectively unchanged for this reprocessing, with one exception. For MODIS, the OBPG has previously been distributing Level-1A files with non-ocean bands removed. This is referred to as band-subsetted Level-1A. In this reprocessing, the OBPG will restore the Level-1A archive to the full-band-suite format. A typical Level-1A file will therefore increase in size from 215MB to 575MB, with compressed files typically increasing from 50MB to 220MB. The disadvantage of this change is obvious: files are bigger and transfers will take more time. The advantages of this change are many:

We recognize that some users will find this file-size increase difficult to manage. Thus, we plan to add support for band-subsetting in the OBPG ordering system, but direct access distribution mechanisms will be delivering the full-band-suite Level-1A. This transition will occur with the reprocessing, so there will be a period of time when MODIS-Aqua is full-band and MODIS-Terra is subsetted.

Level-2 Product Content

The standard Level-2 ocean color (OC) product suite has been expanded to include additional derived products. The epsilon field has been removed, and the water-leaving radiance (nLw) fields have been replaced with remote sensing reflectances (Rrs). It is currently planned that the MODIS products will also include Rrs from the land channels at 469, 555, and 645-nm, aggregated to the nominal 1-km resolution of the ocean color channels. The MODIS ocean color band at 551 is now reported at 547-nm, as this is consistent with the true center of the bandpass and thus reduces the need for model-based band-pass corrections. Changes have also been made to Level-2 geolocation and meta-data as noted below. The fundamental format is still hdf4.

  1. Science datasets included in a standard SeaWiFS Level-2 OC file

    • aot_865, aerosol optical thickness at 865-nm
    • angstrom, aerosol Angstrom exponent at 443-nm relative to 865-nm
    • Rrs_412, remote sensing reflectance at 412-nm
    • Rrs_443, remote sensing reflectance at 443-nm
    • Rrs_490, remote sensing reflectance at 490-nm
    • Rrs_510, remote sensing reflectance at 510-nm
    • Rrs_555, remote sensing reflectance at 555-nm
    • Rrs_670, remote sensing reflectance at 670-nm
    • chlor_a, chlorophyll-a (OC4v5 algorithm)
    • Kd_490, diffuse attenuation at 490-nm (KD2 algorithm)
    • pic, particulate inorganic carbon (Gordon and Balch algorithm)
    • poc, particulate organic carbon (Stramski algorithm)
    • cdom_index, colored disolved organic matter index (Morel algorithm)
    • par, daily mean photosynthetically available radiation (Frouin algorithm)
  2. Science datasets included in a standard MODIS Level-2 OC file

    • aot_869, aerosol optical thickness at 869-nm
    • angstrom, aerosol Angstrom exponent at 443-nm relative to 869-nm
    • Rrs_412, remote sensing reflectance at 412-nm
    • Rrs_443, remote sensing reflectance at 443-nm
    • Rrs_469, remote sensing reflectance at 469-nm
    • Rrs_488, remote sensing reflectance at 488-nm
    • Rrs_531, remote sensing reflectance at 531-nm
    • Rrs_547, remote sensing reflectance at 547-nm
    • Rrs_555, remote sensing reflectance at 555-nm
    • Rrs_645, remote sensing reflectance at 645-nm
    • Rrs_667, remote sensing reflectance at 667-nm
    • Rrs_678, remote sensing reflectance at 678-nm
    • chlor_a, chlorophyll-a (OC3v5 algorithm
    • )
    • Kd_490, diffuse attenuation at 490-nm (KD2 algorithm)
    • pic, particulate inorganic carbon (Gordon and Balch algorithm)
    • poc, particulate organic carbon (Stramski algorithm)
    • cdom_index, colored disolved organic matter index (Morel algorithm)
    • par, daily mean photosynthetically available radiation (Frouin algorithm)
    • ipar, instantaneous photosynthetically available radiation
    • nflh, fluorescence line height (normalized)
  3. Geo-location data

    • The longitude and latitude datasets within the Level-2 products were previously provided in a sub-sampled form. They will now be full resolution. This simplifies usage by non-SeaDAS applications.
  4. Fill values and Meta Data

    • The fill-values have been standardized across datasets to simplify the identification of masked or failed retrievals. In most cases, the derived product fields are now filled with -32767. Meta-data attributes were added to each field to indicate the fill value both before and after any scaling to integer. In the case where datasets are stored as scaled integers, the fill values are provided for both the scaled and unscaled values.For the Rrs datasets, the solar irradiance is also included in the meta-data (in units of mW/cm^2/um/sr) to simplify conversion back to nLw (i.e., nLw = Rrs*solar_irradiance).

    e.g.:

    int16 Rrs_670(Number of Scan Lines, Pixels per Scan Line) ;
    Rrs_670:long_name = "Remote sensing reflectance at 670 nm" ;
    Rrs_670:slope = 2.e-06f ;
    Rrs_670:intercept = 0.050000001f ;
    Rrs_670:units = "sr^-1" ;
    Rrs_670:solar_irradiance = 151.22081f ;
    Rrs_670:bad_value_scaled = -32767s ;
    Rrs_670:bad_value_unscaled = -0.015534f ;

    A processing version identifier was also added to the global attributes. For this reprocessing it the "pversion" attribute will be set to 2009.1.

  5. File naming convention

    • Level-2 file names will now include the suite identifier (OC in this case). For example, a SeaWiFS Level-2 filename would be of the form Syyyydddhhmmss.L2_GAC_OC. This is to accomodate future product suites such as inherent optical properties (IOP).

Level-3 Product Content

The Level-3 binned file format used by the OBPG includes one or more geophysical parameters and one set of statistics (i.e., number of observations, sum of weights, etc.) per bin. That means that all parameters included in a given bin file must share the same exclusion criteria (masking). It is therefore desireable to separate the geophysicical products into logical combinations (or suites) with similar retrieval-failure rates.

For this reprocessing, the standard Level-2 OC products will be composited into a number of such logical groupings. The suite names and corresponding geophysical parameters are listed for each sensor below. The binned file names will now include the suite name (e.g., Syyyyddd.L3b_DAY_RRS.main).

  1. Standard SeaWiFS Level-3 binned files

    RRS suite
    aot_865aerosol optical thickness at 865-nm
    angstromaerosol Angstrom exponent at 443-nm relative to 865-nm
    Rrs_412remote sensing reflectance at 412-nm
    Rrs_443remote sensing reflectance at 443-nm
    Rrs_490remote sensing reflectance at 490-nm
    Rrs_510remote sensing reflectance at 510-nm
    Rrs_555remote sensing reflectance at 555-nm
    Rrs_670remote sensing reflectance at 670-nm
    CHL suite
    chlor_achlorophyll-a (OC4v5 algorithm)
    KD490 suite
    Kd_490diffuse attenuation at 490-nm (KD2 algorithm)
    PIC suite
    picparticulate inorganic carbon (Gordon and Balch algorithm)
    POC suite
    pocparticulate organic carbon (Stramski algorithm)
    CDOM suite
    cdom_indexcolored disolved organic matter index (Morel algorithm)
    PAR suite
    pardaily mean photosynthetically available radiation (Frouin algorithm)

    Sample Files Available Here

  2. Standard MODIS Level-3 binned files

    RRS suite
    aot_869aerosol optical thickness at 869-nm
    angstromaerosol Angstrom exponent at 443-nm relative to 869-nm
    Rrs_412remote sensing reflectance at 412-nm
    Rrs_443remote sensing reflectance at 443-nm
    Rrs_469remote sensing reflectance at 469-nm
    Rrs_488remote sensing reflectance at 488-nm
    Rrs_531remote sensing reflectance at 531-nm
    Rrs_547remote sensing reflectance at 547-nm
    Rrs_555remote sensing reflectance at 555-nm
    Rrs_645remote sensing reflectance at 645-nm
    Rrs_667remote sensing reflectance at 667-nm
    Rrs_678remote sensing reflectance at 678-nm
    CHL suite
    chlor_achlorophyll-a (OC3v5 algorithm)
    KD490 suite
    Kd_490diffuse attenuation at 490-nm (KD2 algorithm)
    PIC suite
    picparticulate inorganic carbon (Gordon and Balch algorithm)
    POC suite
    pocparticulate organic carbon (Stramski algorithm)
    CDOM suite
    cdom_indexcolored disolved organic matter index (Morel algorithm)
    PAR suite
    pardaily mean photosynthetically available radiation (Frouin algorithm)
    FLH suite
    iparinstantaneous photosynthetically available radiation
    nflhfluorescence line height (normalized)
  3. Standard Mapped Image (SMI) files

    The standard Level-3 mapped file format has been changed from a gridded 16-bit scaled integer array to a floating-point array. This was done to increase the science value of the mapped datasets, by delaying arbitrary scaling decisions until the point of actually generating a viewable image. In float format, the SMI product is effectively just a mapped version of the binned product. The SMI meta-data has also been changed to better support the float format. The global and SDS attributes of scaling, slope, and intercept are now set to "linear", 1.0, and 0.0, respectively, to ensure that software that blindly applies these values to "unscale" the mapped data (l3m_data SDS) back to float will have no effect. e.g.:

    Parameter = "Chlorophyll a concentration" ;
    Measure = "Mean" ;
    Units = "mg m^-3" ;
    Scaling = "linear" ;
    Scaling Equation = "(Slope*l3m_data) + Intercept = Parameter value" ;
    Slope = 1.f ;
    Intercept = 0.f ;

    Additional meta-data was added to provide a suggested scaling range that can be used by downstream applications to convert the float image into a byte-scaled image for display. This is the information that the OBPG uses to generate the Level-3 browse PNG images. e.g.:

    Suggested Image Scaling Minimum = 0.01f ;
    Suggested Image Scaling Maximum = 20.f ;
    Suggested Image Scaling Type = "LOG" ;
    Suggested Image Scaling Applied = "No" ;

    The naming convention has also changed to capture the Level-3 product suite name and the approximate resolution at the equator. Standard resolutions are still 9km and 4km for SeaWiFS and MODIS, respectively. Projection is still equirectangular.As an example, a full set of SeaWiFS mapped files for one day (2005 day 264) would look like this:

    • S2005264.L3m_DAY_RRS_angstrom_9km
    • S2005264.L3m_DAY_RRS_aot_865_9km
    • S2005264.L3m_DAY_RRS_Rrs_412_9km
    • S2005264.L3m_DAY_RRS_Rrs_443_9km
    • S2005264.L3m_DAY_RRS_Rrs_490_9km
    • S2005264.L3m_DAY_RRS_Rrs_510_9km
    • S2005264.L3m_DAY_RRS_Rrs_555_9km
    • S2005264.L3m_DAY_RRS_Rrs_670_9km
    • S2005264.L3m_DAY_CHL_chlor_a_9km
    • S2005264.L3m_DAY_KD490_Kd_490_9km
    • S2005264.L3m_DAY_POC_poc_9km
    • S2005264.L3m_DAY_PIC_pic_9km
    • S2005264.L3m_DAY_PAR_par_9km
    • S2005264.L3m_DAY_CDOM_cdom_index_9km