MyOcean Surface Chlorophyll
Provider: MyOcean (CNR)
Contact person: ???
Contact for production: MyOcean Service Desk Send mail
Algorithm development: ICESS
Chl-a is used as an estimate of phytoplankton abundance or biomass though chl-a is a pigment and the ratio of chl-a to carbon in a phytoplankton assemblage is quite variable. Chl-a can be used to compute primary production and in high concentrations can indicate eutrophication. Remotely sensed Chl-a is used by international, national and regional research institutions, and environmental organizations for change detection, time series analysis, early warning of harmful algal blooms, and productivity. Data are produced for the Mediterranean by the Group for Satellite Oceanography (GOS-ISAC) of the Italian National Research Council (CNR), in Rome, and globally by ACRI.
Units: Milligrams of chl-a per m3 (mg m-3); values range from ~0.01 to 64
Data Portal: MyOcean 2
Platform & processing
The Group for Satellite Oceanography (GOS-ISAC) collects Level-0 raw data from ESA and/or NASA as soon as they are available an classed as Near Real Time (NRT). Delayed Time (DT) processing mode is performed four days after satellite overpasses as soon as ancillary data are made available for downloading from NASA. Standard masking criteria for detecting clouds or other contamination factors are routinely applied, i.e., land, cloud, sun glint, atmospheric correction failure, high total radiance, large solar zenith angle (70deg.), large spacecraft zenith angle (56deg.), coccolithophores, negative water leaving radiance, and normalized water leaving radiance at 555 nm.[McClain et al., 1995].
Global Ocean, European Seas / North Africa
The ESA Envisat MERIS radiance products provided by ACRI are the outputs of the case 1 / open ocean or case 2 / coastal water algorithms at 1.2 km resolution from 2002 to date. The NASA radiances are case 1: Aqua MODIS at ~1.1 km resolution and SeaWiFS at 2km resolution. ESA data are processed by ESA to level 2 using specific software (see http://www.odesa-info.eu/distrib/); NASA data are processed to level 2 using NASA specific software. The user version of the processing software is called SEADAS and is available at http://oceancolor.gsfc.nasa.gov/seadas. SeaDAS can also process MERIS data.
MODIS data for the Mediterranean area are processed using the Mediterranean regional Ocean Color algorithm (MedOC3, Santoleri et al., 2008). This algorithm was developed by the GOS-ISAC and is an empirical ocean color algorithm for chlorophyll retrieval most suited for the Mediterranean Case 1 waters. For MERIS the chlorophyll product is obtained by means of the MedOC4Me algorithm (Mediterranean Ocean Color 4 bands, Santoleri et al., 2008). MedOC4Me is also an empirical ocean color algorithm for chlorophyll retrieval. It uses the blue-to-green Reflectance ratio. In particular, it uses, on a pixel-by-pixel basis, the Maximum among Rrs443/Rrs560, Rrs490/Rrs560 and Rrs510/Rrs560, where Rrs443, Rrs490, Rrs510 and Rrs560 are the Remote Sensing Reflectances at 443, 490, 510 and 560 nm, respectively. CHL product identifies the average chlorophyll content of the surface layer as defined by the first optical depth (roughly one fifth of the euphotic depth).
Error estimates suggest that the GSM chlorophyll-a product accuracy is of the order of 10-0.28, error estimates for the single sensor components were: 10-0.28 for SeaWiFS, 10-0.35 for MODIS, and 10-0.30 for MERIS (http://www.globcolour.info/validation/report/GlobCOLOUR_FVR_v1.1.pdf). Individual per pixel error estimates, determined from the optimisation, are estimated and included as are the processing flags.
Ocean color for each cell above the ocean.
Data are only available over the ocean. Information is made available for each product; data quality is made available to the users within the products.
Policy on versioning is unclear and information does not appear to be well publicised. Current version: 2.0b1. Version number embedded in attributes in netCDF files.
The MyOcean sourced data are distributed via the internet.
Typical delay of near-real-time product is a minimum of 24hrs from acquisition of data by ACRI, and the delayed time product is 31 days.
The algorithms to produce chl-a and the atmospheric correction are peer-reviewed
Initial algorithm tuning for various parameters was carried out with in-situ data and validation with in-situ normalised water leaving radiances was also performed (Maritorena et al. 2002). GlobColour used initial matchups with GlobColour chlorophyll primarily to initialise the sensor specific uncertainties which are used in the GSM model. Whether this is an ongoing process is unclear.
Comparison between GlobColour, REASoN and OBPG product sets The Research, Education, and Applications Solution Network (REASoN) project for ocean colour was initiated in 2003 and one of its objectives was to develop an ocean colour time-series from NASA's current sensors (SeaWiFS, MODIS-Aqua) and possibly from the past historical sensors (CZCS, OCTS). One component of Ocean Colour REASoN is to generate and distribute unified data products from the merging of the SeaWiFS and MODIS-AQUA data. This merging activity is conducted by the Institute for Computational Earth System Science (ICESS) at the University of California at Santa Barbara (UCSB). This group developed the GSM semi-analytical ocean colour model (Maritorena et al., 2002) and has shown that it can be used to merge satellite data at the normalized water-leaving radiance level to retrieve several different ocean colour products, namely CHL, CDOM and BBP, simultaneously (Maritorena & Siegel, 2005).
Within the REASoN project, the GSM model is routinely used to merge the data from SeaWiFS and MODIS-AQUA. In REASoN, the GSM products are derived for the complete archive of each individual sensor (9/1997 to today for SeaWiFS, 7/2002 to today for AQUA) and from the merged Lwn(λ) data for the period common to both sensors (7/2002 to today). These products, along with their confidence intervals are accessible to public users as mapped (4320 columns X 2160 rows) HDF files through an anonymous ftp server at: ftp://ftp.oceancolor.ucsb.edu//pub/org/oceancolor/REASoN. [ SeaWiFS/MODIS-AQUA, GSM, chl, cdom, bbp, uncertainties, 9km, daily, 4day, 8day, monthly]
A SeaWiFS-MODIS-Aqua merged CHL product is also available through the NASA Ocean Biology Processing Group (OBPG) at Goddard Space Flight Centre (GSFC). This merged chlorophyll product is calculated using a weighted average scheme similar to what is used for the temporal binning of SeaWiFS and MODIS data (Campbell et al., 1995). The OBPG merged CHL is available for daily, 8-day, monthly, seasonal and annual time periods . [ SeaWiFS/MODIS-Aqua, weighted average, chl, 9km, daily, 8day, monthly, seasonally, yearly]
The observed differences between the GSM-derived chlorophyll and those from OBPG are discussed in Siegel et al. (2005).
The biggest differences are observed when comparing the results from the individual sensors. For example, when MERIS is used alone (GlobColour), its chlorophyll retrievals tend to be slightly higher than those from SeaWiFS or Aqua. The opposite is frequently true for MODIS-Aqua which often shows the lowest chlorophyll values. The MERIS high values influence the GlobColour merged product which tends to be higher than the other two in areas where the MERIS data alone return high values compared to the other sensors. Coastal zones (depth < 1000 m) and high latitudes are areas where big differences are sometimes observed between the GlobColour, REASoN and OBPG products. Large differences (> 25%) in the chlorophyll products are also observed for areas where dust and aerosols can be present in significant amounts (e.g. Mediterranean and Central Atlantic). In some areas, a seasonal trend appears in the differences between products with larger differences being observed at peak chlorophyll values, when MERIS is involved in particular. Areas with marked seasonal variations often show the smallest differences among data sets for low values of the seasonal cycle. For waters deeper than 1000m and within the 50°N- 50°S latitudinal range, the merged chlorophyll product from GlobColour or REASoN agree within 10% or less with the merged CHL product from the OBPG. The GlobColour and REASoN merged chlorophylls agree within 15% or less for the same waters. Over the 30 areas considered, the chlorophyll merged data sets agree within ~15% on average. What is of concern is that the single-sensor 'merged' products do not agree.
GlobColour carried out a validation exercise which included matchups (approx. 700) with in-situ data of chlorophyll and water leaving radiances (http://www.globcolour.info/validation/report/GlobCOLOUR_FVR_v1.1.pdf). The main issue is the quality of the chlorophyll estimates produced from the individual sensors data sets: the MERIS derived chlorophyll tends to overestimate, and MODIS-Aqua and SeaWiFS underestimate. With good coverage from all 3 sensors, the sensors compensate for each other, whereas the discrepancy is at its greatest when MERIS isn't included, or is on its own.
Map projection: cylindrical equirectangular
European Seas / North Africa
Map projection: regular equirectangular
Resolution: 4km, 25km or 100km resolution.
Map projection: All resolutions now use regular equirectangular.
The typical delay of product is less than 24 hours for Near Real Time (NRT) products. Delayed Time (DT) products are made available to the user with 5 days of acquisition. DT data are processed using precision orbital data and improved meteorological fields used for atmospheric correction when compared to NRT data. Reanalysis (RAN) products for the Mediterranean are the output of period but infrequent (after some years) reprocessings using the same configuration throughout. Global RAN products are available from 30 days after acquisition as daily, 8-day and monthly composites
Information about processing version changes are not widely disseminated. If processing changes the archive should usually be reprocessed for consistency with current NRT data although whether this mechanism is in place is not known.For some datasets the RAN products are continuously updated 30 days after satellite acquisition, but in this case there is no guarantee that the configuration is unchanged.
Garver, S. A., & Siegel, D. A. (1997). Inherent optical property inversion of ocean color spectra and its biogeochemical interpretation: I. Time series from the Sargasso Sea. Journal of Geophysical Research, 102, 18607 - 18625.
Gordon, H. R., Brown, O. B., Evans, R. H., Brown, J. W., Smith, R. C., Baker, K. S., et al. (1988). A semi-analytic radiance model of ocean color. Journal of Geophysical Research, 93, 10909 - 10924.
Maritorena, S., Siegel, D. A., & Peterson, A. (2002). Optimization of a semi-analytical ocean color model for global scale applications. Applied Optics, 41(15), 2705 - 2714
Maritorena, S., Siegel, D. A. (2005). Consistent merging of satellite ocean color data sets using a bio-optical model. Remote Sensing of Environment 94:429-440. [pdf]
Santoleri, R., et al. (2008), Open Waters Optical Remote Sensing of the Mediterranean Sea, in Remote Sensing of the European Seas, edited by V. Barale and M. Gade, pp. 103-116, Springer.
Siegel, D. A. S. Maritorena, N. B. Nelson, M. J. Behrenfeld, and C. R. McClain. 2005. Colored dissolved organic matter and its influence on the satellite-based characterization of the ocean biosphere. Geophys. Res. Letters, 32, L20605, doi:10.1029/2005GL024310.