Reprocessing 2014.0 is a multi-mission reprocessing to implement community updates to the ocean color product suite and existing product algorithms, and to incorporate new knowledge of sensor-specific instrument c alibration. Note that this is the first reprocessing since 2009-2010 in which significant changes will be made to product algorithms and content. This 2014.0 reprocessing will also include a major format transition from HDF4 to netCDF4 for all Level-2 and Level-3 products. This format transition is needed to ensure long-term maintainability and expandability of the derived products produced by the NASA/OBPG, and enhance interoperability with third-party software tools and oce an color (and sea surface temperature) products from other agencies and projects (e.g., ESA, GHRSST).
The instrument calibration will be updated for all sensors to include latest knowledge of sensor degradation and radiometric performance. Significant changes are likely for some sensors, including SeaWiFS, MODIS/Terra and VIIRS. TBD.
The vicarious calibration will be updated for all sensors. For sensors that are calibrated to MOBY (e.g, SeaWiFS and later missions), the update will include extension of the MOBY time-series, which was previously limited to 2007 and prior years.
For the 2014.0 reprocessing, the atmospheric correction algorithm will remain effectively unchanged from the previous processing. Changes to calibration and ancillary sources, however, will impact remote sensing reflectance, Rrs, retrievals and trends.
A significant change has been implemented for the default chlorophyll algorithm, chlor_a, wherein the current standard OC3/OC4 (OCx) band ratio algorithm has been merged with the color index (C I) of Hu et al. (2012). As described in the paper, this refinement is restricted to relatively clear water, and the general impact is to reduce artifacts and biases in clear-water chlorophyll retrievals due to residual glint, stray light, a tmospheric correction errors, and white or spectrally-linear bias errors in Rrs. As implemented, the algorithm diverges slightly from what was published in Hu et al. (2012) in that the transition between CI and OCx now occurs at 0.15 < CI < 0.2 mg/m^3 to ensure a smooth transition. Due to concerns raised within the community with respect to this algorithm change, we will also continue to produce the standard OCx chlorophyll product to support community evaluation. This legacy chlorophyll product will be available in the Level-2 and Level-3 products as chl_ocx, in parallel with the updated chlor_a product. See also presentation by Fr anz from the 2014 Ocean Color Research Team meeting.
The particulate inorganic carbon (pic) algorithm was updated to include a number of refinements from the algorithm PI based on additional field measurements. See presentation by Balch from the 2014 MODIS Science Team Meeting.
The set of standard products will be expanded in this reprocessing to inc lude a full suite of inherent optical properties based on the community-developed GIOP model in the default configuration as described in Werdell et al. (2013). Also based on community feedback, this will include the total absorption and to tal backscatter coefficients at all visible wavelengths, as well as component absorption and scattering coefficients at 443nm with associated spectral slopes and uncertainties. See also presentation by Werdell from the 2014 MODIS Science Team Meeting.
In addition, the IOP model and atmospheric correction algorithm (NIR iteration scheme) will now utilize the backscattering coefficients for pure seawater as developed by Zhang et al. (2009), including temperature and salinity dependence. Previous reprocessing relied on Smith and Baker (1981). Werdell et al. (2013b) recently demonstrated the significance of this change in the context of IOP model retrievals. The temperature and salinity inputs to the Zhang model will be taken from Reynolds 0.25-deg optimally interpolated SST and World Ocean Atlas salinity climatology, respectively.
Since we will now have a standard IOP product suite that will include absorption due to gelbstoff and detrital material, the CDOM_index product will be discontinued as a standard product. Other existing product algorithms will remain effectively unchanged, with only minor corrections and refinements.
Standard ocean color processing relies on a number of ancillary data streams, e.g., meteorological conditions (atmospheric pressure, wind speeds, relative humidity, water vapor concentration), absorbing atmospheric gas concentrations (ozone, NO2), and seaice distribution.
The current standard meteorological data source is the National Center for Environmental Prediction (NCEP). For this reprocessing, the NCEP data will be updated to utilize improvements from the NCEP-DOE 2 Reanalysis. In effect, the existing NCEP record will be renormalized to track the NCEP-DOE 2 reanalysis. The renormalization approach retains the higher spatial resolution of the existing NCEP record (1x1-deg) while taking advantage of the improved temporal consistency of the reanalysis (2.5x2.5-deg). In previous reprocessing, this normalization process was only performed for the relative humidity field. The impact of these ancillary changes is expected to be minor. A study of the changes of the new meteorological data is in A comparison of the new meteorological data to the currently used meteorological data. The meteorological data climatology was also replaced. Until this reprocessing, the climatology was derived from the COADS (Comprehensive Ocean-Atmosphere Data Set) data. The current monthly climatology has been derived from the daily meteorological data from 1997 to 2014. It has better coverage and higher resolution (from 2.5 to 1 degree resolution). See a comparison of the 2 climatologies here.
The standard ozone data source is a concatenation of multiple missions and sensors including N7/TOMS, EP/TOMS, AURA/OMI. To mitigate sensor-to-sensor biases, the multi-mission ozone time-series will be renormalized to track the long-term SBUV climate record (Labow et al. 2013). As with the NCEP update, the ozone update is designed to retain the higher spatial resolution of the TOMS-OMI record while taking advantage of the more temporally-consistent SBUV record. The impact of this change is up to a few percent in Rrs retrievals in the red and green, with much smaller effects in other wavelengths (e.g., see impact to SeaWiFS Rrs time-series). The ozone climatology was also improved by using a ten year span of daily ozone files to make the daily climatology.
The NO2 data source has been updated. Prior to the 2014 reprocessing, the processing used a monthly climatology of NO2 derived from 4 years of NO2 daily data from the OMI instrument aboard the AURA spacecraft. That climatology has been updated to use 8 years of AURA OMI data.
NOAA's ETOPO1 Global Relief Model is used for both land height and water depth. The Land/Water mask was generated using Generic Mapping Tools (GMT), from the GSHHG 2.3.2 data set.
Ancillary data sources are further described on the Ancillary Data Sources page.
For each Level-1 s atellite granule, two Level-2 files will be produced, each containing a different set of derived products. These are referred to as the ocean color (OC) suite and the inherent optical properties (IOP) suite. The OC suite will be very similar to the currently produced OC suite.
Level-2 OC Suite
Level-2 IOP Suite
The Level-3 binning scheme is unchanged from previo us reprocessings. Each Level-2 product will be spatially and temporally composited into a Level-3 binned (L3b) product at approximately 4.6x4.6 km2 or 9.2x9.2 km2, depending on sensor native resolution. The bin files w ill be grouped into several suites based on common flagging and masking requirements and shared algorithm retrieval performance. This will include:
Level-3 RRS Suite
Level-3 CHL Suite
Level-3 KD490 Suite
Level-3 PIC Suite
Level-3 POC Suite
Level-3 PAR Suite
Level-3 FLH Suite (MODIS only)
Level-3 IOP Suite
Finally, every Level-3 product in each suite will separately mapped to a standard equi-rectangular gridded (L3m) product. The mapping scheme is unchanged from previous reprocessings.
A primary motivation for this reprocessing is to adopt a more efficient, sustainable, extendable, and interoperable data file format. Specifically, all Level-2 and Level-3 files produced and distributed by the Ocean Biology Processing Group for ocean color and SST will transition from a base format of HDF4 to netCDF4 and will utilize a variety of international standards and conventions for meta-data and file structure. A detailed description of the meta-data standards and links to sample files is available here:
File format and meta-data changes.
In the past, the production of Refined Level-2 products was limited by the availability of refined attitude and ephemeris data, which typically were produced within four days of the initial "Near Real Time" predicted products. With the transition to NCEP-DOE 2 Reanalysis ancillary meteorological data, which become available monthly on the 15th day of the next month, the latency of Refined Level-2 products will increase to 15 - 45 days. Our Near Real Time "Quick Look" (QL) evaluation products will continue to use NCEP Global Data Analysis System (GDAS) meteorological data which are usually available about 7 hours after they are collected. The QL and Refined data can be identified by the 'pversion' attribute in the 'input_parameters' group in the file metadata:
QL:
group: input_parameters {
// group attributes:
:pversion = "2014.0QL" ;
Refined:
group: input_parameters {
// group attributes:
:pversion = "2014.0" ;
Comparison of results relative previous reprocessing and relative to in situ measurements (where available) will be provided for each sensor prior to initiation of reprocessing for that sensor.
